diff --git a/sycl/include/sycl/ext/intel/esimd/memory_properties.hpp b/sycl/include/sycl/ext/intel/esimd/memory_properties.hpp index b25f24bdc829c..94b3b2cebd61d 100644 --- a/sycl/include/sycl/ext/intel/esimd/memory_properties.hpp +++ b/sycl/include/sycl/ext/intel/esimd/memory_properties.hpp @@ -238,13 +238,17 @@ template <typename PropertyListT> struct remove_alignment_property { }; template <size_t Alignment, typename... LastTs> struct remove_alignment_property< - properties<std::tuple<alignment_key::value_t<Alignment>, LastTs...>>> { - using type = properties<std::tuple<LastTs...>>; + properties<oneapi::experimental::detail::properties_type_list< + alignment_key::value_t<Alignment>, LastTs...>>> { + using type = + properties<oneapi::experimental::detail::properties_type_list<LastTs...>>; }; template <typename FirstT, size_t Alignment, typename... LastTs> -struct remove_alignment_property<properties< - std::tuple<FirstT, alignment_key::value_t<Alignment>, LastTs...>>> { - using type = properties<std::tuple<FirstT, LastTs...>>; +struct remove_alignment_property< + properties<oneapi::experimental::detail::properties_type_list< + FirstT, alignment_key::value_t<Alignment>, LastTs...>>> { + using type = properties< + oneapi::experimental::detail::properties_type_list<FirstT, LastTs...>>; }; template <typename PropertyListT> using remove_alignment_property_t = @@ -323,8 +327,11 @@ struct property_value<__ESIMD_NS::cache_hint_L3_key, // Declare that esimd::properties is a property_list. template <typename... PropertyValueTs> -struct is_property_list<__ESIMD_NS::properties<std::tuple<PropertyValueTs...>>> - : is_property_list<properties<std::tuple<PropertyValueTs...>>> {}; +struct is_property_list<__ESIMD_NS::properties< + oneapi::experimental::detail::properties_type_list<PropertyValueTs...>>> + : is_property_list< + properties<oneapi::experimental::detail::properties_type_list< + PropertyValueTs...>>> {}; namespace detail { // We do not override the class ConflictingProperties for cache_hint properties diff --git a/sycl/include/sycl/ext/oneapi/experimental/annotated_arg/annotated_arg.hpp b/sycl/include/sycl/ext/oneapi/experimental/annotated_arg/annotated_arg.hpp index cea0f6ddd745f..91baca3e14e3f 100644 --- a/sycl/include/sycl/ext/oneapi/experimental/annotated_arg/annotated_arg.hpp +++ b/sycl/include/sycl/ext/oneapi/experimental/annotated_arg/annotated_arg.hpp @@ -66,7 +66,8 @@ annotated_arg(T, Args...) -> annotated_arg<T, typename detail::DeducedProperties<Args...>::type>; template <typename T, typename old, typename... ArgT> -annotated_arg(annotated_arg<T, old>, properties<std::tuple<ArgT...>>) +annotated_arg(annotated_arg<T, old>, + properties<detail::properties_type_list<ArgT...>>) -> annotated_arg< T, detail::merged_properties_t<old, detail::properties_t<ArgT...>>>; diff --git a/sycl/include/sycl/ext/oneapi/experimental/annotated_ptr/annotated_ptr.hpp b/sycl/include/sycl/ext/oneapi/experimental/annotated_ptr/annotated_ptr.hpp index 21d8e8333964e..ea90f3d9d1cef 100644 --- a/sycl/include/sycl/ext/oneapi/experimental/annotated_ptr/annotated_ptr.hpp +++ b/sycl/include/sycl/ext/oneapi/experimental/annotated_ptr/annotated_ptr.hpp @@ -61,8 +61,9 @@ using PropertiesFilter = // filter properties that are applied on annotations template <typename... Props> -using annotation_filter = properties< - PropertiesFilter<std::tuple<Props...>, propagateToPtrAnnotation>>; +using annotation_filter = + properties<PropertiesFilter<detail::properties_type_list<Props...>, + propagateToPtrAnnotation>>; } // namespace detail template <typename I, typename P> struct annotationHelper {}; @@ -245,7 +246,8 @@ annotated_ptr(T *, Args...) -> annotated_ptr<T, typename detail::DeducedProperties<Args...>::type>; template <typename T, typename old, typename... ArgT> -annotated_ptr(annotated_ptr<T, old>, properties<std::tuple<ArgT...>>) +annotated_ptr(annotated_ptr<T, old>, + properties<detail::properties_type_list<ArgT...>>) -> annotated_ptr< T, detail::merged_properties_t<old, detail::properties_t<ArgT...>>>; #endif // __cpp_deduction_guides diff --git a/sycl/include/sycl/ext/oneapi/experimental/work_group_memory.hpp b/sycl/include/sycl/ext/oneapi/experimental/work_group_memory.hpp index 7870ebd3ca73e..8ffab28b965a3 100644 --- a/sycl/include/sycl/ext/oneapi/experimental/work_group_memory.hpp +++ b/sycl/include/sycl/ext/oneapi/experimental/work_group_memory.hpp @@ -77,7 +77,7 @@ class __SYCL_SPECIAL_CLASS __SYCL_TYPE(work_group_memory) work_group_memory void __init(decoratedPtr ptr) { this->ptr = ptr; } #endif private: - decoratedPtr ptr; + decoratedPtr ptr = nullptr; }; } // namespace ext::oneapi::experimental } // namespace _V1 diff --git a/sycl/include/sycl/ext/oneapi/properties/properties.hpp b/sycl/include/sycl/ext/oneapi/properties/properties.hpp index b7db9715cad4d..cecb310b97eb7 100644 --- a/sycl/include/sycl/ext/oneapi/properties/properties.hpp +++ b/sycl/include/sycl/ext/oneapi/properties/properties.hpp @@ -69,87 +69,6 @@ static constexpr std::enable_if_t<!HasProperty, void> get_property() { return; } -// Filters for all runtime properties with data in a tuple of properties. -// NOTE: We only need storage for runtime properties with data. -template <typename T> struct RuntimePropertyStorage {}; -template <typename... Ts> struct RuntimePropertyStorage<std::tuple<Ts...>> { - using type = std::tuple<>; -}; -template <typename T, typename... Ts> -struct RuntimePropertyStorage<std::tuple<T, Ts...>> - : std::conditional_t<IsRuntimeProperty<T>::value, - PrependTuple<T, typename RuntimePropertyStorage< - std::tuple<Ts...>>::type>, - RuntimePropertyStorage<std::tuple<Ts...>>> {}; - -// Count occurrences of a type in a tuple. -template <typename T, typename Tuple> struct CountTypeInTuple; -template <typename T, typename... TupleTs> -struct CountTypeInTuple<T, std::tuple<TupleTs...>> - : std::integral_constant< - size_t, (0 + ... + static_cast<size_t>(std::is_same_v<T, TupleTs>))> { -}; - -// Helper for counting the number of properties that are also in PropertyArgsT. -template <typename PropertyArgsT, typename Props> struct CountContainedProps; -template <typename PropertyArgsT> -struct CountContainedProps<PropertyArgsT, std::tuple<>> - : std::integral_constant<size_t, 0> {}; -template <typename PropertyArgsT, typename PropertyT, typename... PropertyTs> -struct CountContainedProps<PropertyArgsT, - std::tuple<PropertyT, PropertyTs...>> { - static constexpr size_t NumOccurrences = - CountTypeInTuple<PropertyT, PropertyArgsT>::value; - static_assert(NumOccurrences <= 1, - "Duplicate occurrences of property in constructor arguments."); - static constexpr size_t value = - CountContainedProps<PropertyArgsT, std::tuple<PropertyTs...>>::value + - NumOccurrences; -}; - -// Helper class to extract a subset of elements from a tuple. -// NOTES: This assumes no duplicate properties and that all properties in the -// struct template argument appear in the tuple passed to Extract. -template <typename PropertyArgsT, typename PropertiesT> -struct ExtractProperties; -template <typename PropertyArgsT> -struct ExtractProperties<PropertyArgsT, std::tuple<>> { - static constexpr std::tuple<> Extract(const PropertyArgsT &) { - return std::tuple<>{}; - } -}; -template <typename PropertyArgsT, typename PropertyT, typename... PropertiesTs> -struct ExtractProperties<PropertyArgsT, - std::tuple<PropertyT, PropertiesTs...>> { - static constexpr std::tuple<PropertyT, PropertiesTs...> - Extract(const PropertyArgsT &PropertyValues) { - // TODO: NumOccurrences and checks should be moved out of the function once - // https://github.com/intel/llvm/issues/13677 has been fixed. - constexpr size_t NumOccurrences = - CountTypeInTuple<PropertyT, PropertyArgsT>::value; - static_assert( - NumOccurrences <= 1, - "Duplicate occurrences of property in constructor arguments."); - static_assert(NumOccurrences == 1 || - std::is_default_constructible_v<PropertyT>, - "Each property in the property list must either be given an " - "argument in the constructor or be default-constructible."); - - auto NextExtractedProperties = - ExtractProperties<PropertyArgsT, std::tuple<PropertiesTs...>>::Extract( - PropertyValues); - - if constexpr (NumOccurrences == 1) { - return std::tuple_cat( - std::tuple<PropertyT>{std::get<PropertyT>(PropertyValues)}, - NextExtractedProperties); - } else { - return std::tuple_cat(std::tuple<PropertyT>{PropertyT{}}, - NextExtractedProperties); - } - } -}; - // Get the value of a property from a property list template <typename PropKey, typename ConstType, typename DefaultPropVal, typename PropertiesT> @@ -168,112 +87,211 @@ struct GetPropertyValueFromPropList<PropKey, ConstType, DefaultPropVal, PropertyMetaInfo<std::remove_const_t<prop_val_t>>::value; }; +template <typename... property_tys> +inline constexpr bool properties_are_unique = []() constexpr { + if constexpr (sizeof...(property_tys) == 0) { + return true; + } else { + const std::array kinds = {PropertyID<property_tys>::value...}; + auto N = kinds.size(); + for (std::size_t i = 0; i < N; ++i) + for (std::size_t j = i + 1; j < N; ++j) + if (kinds[i] == kinds[j]) + return false; + + return true; + } +}(); + +template <typename... property_tys> +inline constexpr bool properties_are_sorted = []() constexpr { + if constexpr (sizeof...(property_tys) == 0) { + return true; + } else { + const std::array kinds = {PropertyID<property_tys>::value...}; + // std::is_sorted isn't constexpr until C++20. + for (std::size_t idx = 1; idx < kinds.size(); ++idx) + if (kinds[idx - 1] >= kinds[idx]) + return false; + return true; + } +}(); + +template <typename... property_tys> +constexpr bool properties_are_valid_for_ctad = []() constexpr { + // Need `if constexpr` to avoid hard error in "unique" check when querying + // property kind if `property_tys` isn't a property. + if constexpr (!((is_property_value_v<property_tys> && ...))) { + return false; + } else if constexpr (!detail::properties_are_unique<property_tys...>) { + return false; + } else { + return true; + } +}(); + +template <typename... property_tys> struct properties_type_list; +template <typename... property_tys> struct invalid_properties_type_list {}; } // namespace detail -template <typename PropertiesT> class properties { - static_assert(detail::IsTuple<PropertiesT>::value, - "Properties must be in a tuple."); - static_assert(detail::AllPropertyValues<PropertiesT>::value, - "Unrecognized property in property list."); - static_assert(detail::IsSorted<PropertiesT>::value, - "Properties in property list are not sorted."); - static_assert(detail::SortedAllUnique<PropertiesT>::value, - "Duplicate properties in property list."); - static_assert(detail::NoConflictingProperties<PropertiesT>::value, - "Conflicting properties in property list."); +template <typename properties_type_list_ty> class __SYCL_EBO properties; + +// Empty property list. +template <> class __SYCL_EBO properties<detail::properties_type_list<>> { + template <typename T> + static constexpr bool empty_properties_list_contains = false; public: - template <typename... PropertyValueTs, - std::enable_if_t<detail::AllPropertyValues< - std::tuple<PropertyValueTs...>>::value, - int> = 0> - constexpr properties(PropertyValueTs... props) - : Storage(detail::ExtractProperties<std::tuple<PropertyValueTs...>, - StorageT>::Extract({props...})) { - // Default-constructible properties do not need to be in the arguments. - // For properties with a storage, default-constructibility is checked in - // ExtractProperties, while those without are so by default. As such, all - // arguments must be a unique property type and must be in PropertiesT. - constexpr size_t NumContainedProps = - detail::CountContainedProps<std::tuple<PropertyValueTs...>, - PropertiesT>::value; - static_assert(NumContainedProps == sizeof...(PropertyValueTs), - "One or more property argument is not a property in the " - "property list."); - // We're in process of refactoring properties infrastructure, make sure that - // any newly added properties use `detail::property_base`! - static_assert( - (std::is_base_of_v<detail::property_tag, PropertyValueTs> && ...)); + template <typename property_key_t> static constexpr bool has_property() { + return false; } - template <typename PropertyT> - static constexpr std::enable_if_t<detail::IsProperty<PropertyT>::value, bool> - has_property() { - return detail::ContainsProperty<PropertyT, PropertiesT>::value; + // Never exists for empty property list, provide this for a better error + // message: + template <typename T> + static std::enable_if_t<empty_properties_list_contains<T>> get_property() {} +}; + +// Base implementation to provide nice user error in case of mis-use. Without it +// an error "base class '<property>' specified more than once as a direct base +// class" is reported prior to static_assert's error. +template <typename... property_tys> +class __SYCL_EBO + properties<detail::invalid_properties_type_list<property_tys...>> { +public: + properties(property_tys...) { + if constexpr (!((is_property_value_v<property_tys> && ...))) { + static_assert(((is_property_value_v<property_tys> && ...)), + "Non-property argument!"); + } else { + // This is a separate specialization to report an error, we can afford + // doing extra work to provide nice error message without sacrificing + // compile time on non-exceptional path. Let's find *a* pair of properties + // that failed the check. Note that there might be multiple duplicate + // names, we're only reporting one instance. Once user addresses that, the + // next pair will be reported. + static constexpr auto conflict = []() constexpr { + const std::array kinds = {detail::PropertyID<property_tys>::value...}; + auto N = kinds.size(); + for (int i = 0; i < N; ++i) + for (int j = i + 1; j < N; ++j) + if (kinds[i] == kinds[j]) + return std::pair{i, j}; + }(); + using first_type = detail::nth_type_t<conflict.first, property_tys...>; + using second_type = detail::nth_type_t<conflict.second, property_tys...>; + if constexpr (std::is_same_v<typename first_type::key_t, + typename second_type::key_t>) { + static_assert(!std::is_same_v<typename first_type::key_t, + typename second_type::key_t>, + "Duplicate properties in property list."); + } else { + static_assert( + detail::PropertyToKind<first_type>::Kind != + detail::PropertyToKind<second_type>::Kind, + "Property Kind collision between different property keys!"); + } + } } +}; - template <typename PropertyT> - typename std::enable_if_t<detail::IsRuntimeProperty<PropertyT>::value && - has_property<PropertyT>(), - PropertyT> - get_property() const { - return std::get<PropertyT>(Storage); +// NOTE: Meta-function to implement CTAD rules isn't allowed to return +// `properties<something>` and it's impossible to return a pack as well. As +// such, we're forced to have an extra level of `detail::properties_type_list` +// for the purpose of providing CTAD rules. +template <typename... property_tys> +class __SYCL_EBO properties<detail::properties_type_list<property_tys...>> + : private property_tys... { + static_assert(detail::properties_are_sorted<property_tys...>, + "Properties must be sorted!"); + static_assert( + detail::NoConflictingProperties<std::tuple<property_tys...>>::value, + "Conflicting properties in property list."); + using property_tys::get_property_impl...; + + template <typename> friend class __SYCL_EBO properties; + + template <typename prop_t> static constexpr bool is_valid_ctor_arg() { + return ((std::is_same_v<prop_t, property_tys> || ...)); } - template <typename PropertyT> - typename std::enable_if_t<detail::IsRuntimeProperty<PropertyT>::value && - !has_property<PropertyT>(), - void> - get_property() const { - static_assert(has_property<PropertyT>(), - "Property list does not contain the requested property."); - return; + template <typename prop_t, typename... unsorted_property_tys> + static constexpr bool can_be_constructed_from() { + return std::is_default_constructible_v<prop_t> || + ((false || ... || std::is_same_v<prop_t, unsorted_property_tys>)); } - template <typename PropertyT> - static constexpr auto get_property( - typename std::enable_if_t<detail::IsCompileTimeProperty<PropertyT>::value> - * = 0) { - static_assert(has_property<PropertyT>(), - "Property list does not contain the requested property."); - return detail::get_property<PropertyT, has_property<PropertyT>(), - PropertiesT>(); + // It's possible it shouldn't be that complicated, but clang doesn't accept + // simpler version: https://godbolt.org/z/oPff4h738, reported upstream at + // https://github.com/llvm/llvm-project/issues/115547. Note that if the + // `decltype(...)` is "inlined" then it has no issues with it, but that's too + // verbose. + struct helper : property_tys... { + using property_tys::get_property_impl...; + }; + template <typename property_key_t> + using prop_t = decltype(std::declval<helper>().get_property_impl( + detail::property_key_tag<property_key_t>{})); + +public: + template < + typename... unsorted_property_tys, + typename = std::enable_if_t< + ((is_valid_ctor_arg<unsorted_property_tys>() && ...))>, + typename = std::enable_if_t< + ((can_be_constructed_from<property_tys, unsorted_property_tys...>() && + ...))>, + typename = std::enable_if_t< + detail::properties_are_unique<unsorted_property_tys...>>> + constexpr properties(unsorted_property_tys... props) + : unsorted_property_tys(props)... {} + + template <typename property_key_t> static constexpr bool has_property() { + return std::is_base_of_v<detail::property_key_tag<property_key_t>, + properties>; } -private: - using StorageT = typename detail::RuntimePropertyStorage<PropertiesT>::type; + // Compile-time property. + template <typename property_key_t> + static constexpr auto + get_property() -> std::enable_if_t<std::is_empty_v<prop_t<property_key_t>>, + prop_t<property_key_t>> { + return prop_t<property_key_t>{}; + } - StorageT Storage; + // Runtime property. + // Extra operand to make MSVC happy as it complains otherwise: + // https://godbolt.org/z/WGqdqrejj + template <typename property_key_t> + constexpr auto get_property(int = 0) const + -> std::enable_if_t<!std::is_empty_v<prop_t<property_key_t>>, + prop_t<property_key_t>> { + return get_property_impl(detail::property_key_tag<property_key_t>{}); + } }; -#ifdef __cpp_deduction_guides // Deduction guides -template <typename... PropertyValueTs> +template <typename... PropertyValueTs, + typename = std::enable_if_t< + detail::properties_are_valid_for_ctad<PropertyValueTs...>>> properties(PropertyValueTs... props) -> properties<typename detail::Sorted<PropertyValueTs...>::type>; -#endif -using empty_properties_t = properties<std::tuple<>>; - -// Property list traits -template <typename propertiesT> struct is_property_list : std::false_type {}; -template <typename... PropertyValueTs> -struct is_property_list<properties<std::tuple<PropertyValueTs...>>> - : std::is_same< - properties<std::tuple<PropertyValueTs...>>, - properties<typename detail::Sorted<PropertyValueTs...>::type>> {}; +template <typename... PropertyValueTs, + typename = std::enable_if_t< + !detail::properties_are_valid_for_ctad<PropertyValueTs...>>> +properties(PropertyValueTs... props) + -> properties<detail::invalid_properties_type_list<PropertyValueTs...>>; -#if __cplusplus > 201402L -template <typename propertiesT> -inline constexpr bool is_property_list_v = is_property_list<propertiesT>::value; -#endif +using empty_properties_t = decltype(properties{}); namespace detail { // Helper for reconstructing a properties type. This assumes that // PropertyValueTs is sorted and contains only valid properties. template <typename... PropertyValueTs> -using properties_t = properties<std::tuple<PropertyValueTs...>>; +using properties_t = + properties<detail::properties_type_list<PropertyValueTs...>>; // Helper for merging two property lists; template <typename LHSPropertiesT, typename RHSPropertiesT> @@ -281,8 +299,9 @@ struct merged_properties; template <typename... LHSPropertiesTs, typename... RHSPropertiesTs> struct merged_properties<properties_t<LHSPropertiesTs...>, properties_t<RHSPropertiesTs...>> { - using type = properties<typename MergeProperties< - std::tuple<LHSPropertiesTs...>, std::tuple<RHSPropertiesTs...>>::type>; + using type = properties< + typename MergeProperties<properties_type_list<LHSPropertiesTs...>, + properties_type_list<RHSPropertiesTs...>>::type>; }; template <typename LHSPropertiesT, typename RHSPropertiesT> using merged_properties_t = @@ -340,9 +359,5 @@ struct all_props_are_keys_of< } // namespace detail } // namespace ext::oneapi::experimental - -template <typename PropertiesT> -struct is_device_copyable<ext::oneapi::experimental::properties<PropertiesT>> - : is_device_copyable<PropertiesT> {}; } // namespace _V1 } // namespace sycl diff --git a/sycl/include/sycl/ext/oneapi/properties/property.hpp b/sycl/include/sycl/ext/oneapi/properties/property.hpp index 41a59337440b8..5b147d93f7e95 100644 --- a/sycl/include/sycl/ext/oneapi/properties/property.hpp +++ b/sycl/include/sycl/ext/oneapi/properties/property.hpp @@ -123,11 +123,21 @@ template <> struct is_property_key_of<foo, SYCL_OBJ> : std::true_type {}; #include <stdint.h> // for uint32_t #include <type_traits> // for false_type +#include <sycl/detail/defines_elementary.hpp> + namespace sycl { inline namespace _V1 { namespace ext { namespace oneapi { namespace experimental { +template <typename> class __SYCL_EBO properties; +// Property list traits +template <typename propertiesT> struct is_property_list : std::false_type {}; +template <typename properties_list_ty> +struct is_property_list<properties<properties_list_ty>> : std::true_type {}; +template <typename propertiesT> +inline constexpr bool is_property_list_v = is_property_list<propertiesT>::value; + namespace detail { // List of all properties. @@ -220,14 +230,27 @@ template <typename PropertyT> struct PropertyToKind { }; struct property_tag {}; + +// This is used to implement has/get_property via inheritance queries. +template <typename property_key_t> struct property_key_tag : property_tag {}; + template <typename property_t, PropKind Kind, typename property_key_t = property_t> -struct property_base : property_tag { +struct property_base : property_key_tag<property_key_t> { using key_t = property_key_t; + +protected: + constexpr property_t get_property_impl(property_key_tag<key_t>) const { + return *static_cast<const property_t *>(this); + } + + // For key_t access in error reporting specialization. + template <typename> friend class __SYCL_EBO properties; + #if !defined(_MSC_VER) // Temporary, to ensure new code matches previous behavior and to catch any - // silly copy-paste mistakes. MSVC can't compile it, but linux-only is enough - // for this temporary check. + // silly copy-paste mistakes. MSVC can't compile it, but linux-only is + // enough for this temporary check. static_assert([]() constexpr { if constexpr (std::is_same_v<property_t, key_t>) // key_t is incomplete at this point for runtime properties. @@ -270,6 +293,7 @@ template <typename PropertyT> struct PropertyID { template <typename PropertyT> struct IsRuntimeProperty : std::bool_constant< + !is_property_list_v<PropertyT> && std::is_base_of_v<property_key_base_tag, PropertyT> && !std::is_base_of_v<compile_time_property_key_base_tag, PropertyT>> {}; @@ -277,6 +301,7 @@ struct IsRuntimeProperty template <typename PropertyT> struct IsCompileTimeProperty : std::bool_constant< + !is_property_list_v<PropertyT> && std::is_base_of_v<property_key_base_tag, PropertyT> && std::is_base_of_v<compile_time_property_key_base_tag, PropertyT>> {}; @@ -301,7 +326,8 @@ template <typename> struct HasCompileTimeEffect : std::false_type {}; template <typename T> struct is_property_key - : std::bool_constant<std::is_base_of_v<detail::property_key_base_tag, T>> { + : std::bool_constant<!is_property_list_v<T> && + std::is_base_of_v<detail::property_key_base_tag, T>> { }; template <typename, typename> struct is_property_key_of : std::false_type {}; diff --git a/sycl/include/sycl/ext/oneapi/properties/property_utils.hpp b/sycl/include/sycl/ext/oneapi/properties/property_utils.hpp index 3280d77aa6258..aee53419498eb 100644 --- a/sycl/include/sycl/ext/oneapi/properties/property_utils.hpp +++ b/sycl/include/sycl/ext/oneapi/properties/property_utils.hpp @@ -24,29 +24,33 @@ namespace sycl { inline namespace _V1 { namespace ext::oneapi::experimental { - -// Forward declaration -template <typename PropertyT, typename... Ts> struct property_value; - namespace detail { +template <typename... property_tys> struct properties_type_list; //****************************************************************************** // Misc //****************************************************************************** -// Checks if a type is a tuple. -template <typename T> struct IsTuple : std::false_type {}; -template <typename... Ts> struct IsTuple<std::tuple<Ts...>> : std::true_type {}; +#if __has_builtin(__type_pack_element) +template <int N, typename... Ts> +using nth_type_t = __type_pack_element<N, Ts...>; +#else +template <int N, typename T, typename... Ts> struct nth_type { + using type = typename nth_type<N - 1, Ts...>::type; +}; + +template <typename T, typename... Ts> struct nth_type<0, T, Ts...> { + using type = T; +}; -// Gets the first type in a parameter pack of types. -template <typename... Ts> -using GetFirstType = typename std::tuple_element<0, std::tuple<Ts...>>::type; +template <int N, typename... Ts> +using nth_type_t = typename nth_type<N, Ts...>::type; +#endif -// Prepends a value to a tuple. -template <typename T, typename Tuple> struct PrependTuple {}; +template <typename T, typename PropList> struct PrependProperty {}; template <typename T, typename... Ts> -struct PrependTuple<T, std::tuple<Ts...>> { - using type = std::tuple<T, Ts...>; +struct PrependProperty<T, properties_type_list<Ts...>> { + using type = properties_type_list<T, Ts...>; }; //****************************************************************************** @@ -71,13 +75,13 @@ struct AllPropertyValues<std::tuple<T, Ts...>> // false the head will be void and the tail will be the full tuple. template <typename T1, bool ShouldSplit> struct HeadSplit {}; template <typename T, typename... Ts> -struct HeadSplit<std::tuple<T, Ts...>, true> { +struct HeadSplit<properties_type_list<T, Ts...>, true> { using htype = T; - using ttype = std::tuple<Ts...>; + using ttype = properties_type_list<Ts...>; }; -template <typename... Ts> struct HeadSplit<std::tuple<Ts...>, false> { +template <typename... Ts> struct HeadSplit<properties_type_list<Ts...>, false> { using htype = void; - using ttype = std::tuple<Ts...>; + using ttype = properties_type_list<Ts...>; }; // Selects the one of two types that is not void. This assumes that at least one @@ -103,30 +107,10 @@ template <typename... Ts> struct Sorted { using sortedProperties = sycl::detail::boost::mp11::mp_sort_q<properties, SortByPropertyId>; using type = - sycl::detail::boost::mp11::mp_rename<sortedProperties, std::tuple>; + sycl::detail::boost::mp11::mp_rename<sortedProperties, + detail::properties_type_list>; }; -// Checks if the types in a tuple are sorted w.r.t. their PropertyID. -template <typename T> struct IsSorted {}; -template <typename... Ts> -struct IsSorted<std::tuple<Ts...>> : std::true_type {}; -template <typename T> struct IsSorted<std::tuple<T>> : std::true_type {}; -template <typename L, typename R, typename... Rest> -struct IsSorted<std::tuple<L, R, Rest...>> - : std::conditional_t<PropertyID<L>::value <= PropertyID<R>::value, - IsSorted<std::tuple<R, Rest...>>, std::false_type> {}; - -// Checks that all types in a sorted tuple have unique PropertyID. -template <typename T> struct SortedAllUnique {}; -template <typename... Ts> -struct SortedAllUnique<std::tuple<Ts...>> : std::true_type {}; -template <typename T> struct SortedAllUnique<std::tuple<T>> : std::true_type {}; -template <typename L, typename R, typename... Rest> -struct SortedAllUnique<std::tuple<L, R, Rest...>> - : std::conditional_t<PropertyID<L>::value != PropertyID<R>::value, - SortedAllUnique<std::tuple<R, Rest...>>, - std::false_type> {}; - //****************************************************************************** // Property merging //****************************************************************************** @@ -136,48 +120,53 @@ struct SortedAllUnique<std::tuple<L, R, Rest...>> // NOTE: This assumes that the properties are in sorted order. template <typename LHSPropertyT, typename RHSPropertyT> struct MergeProperties; -template <> struct MergeProperties<std::tuple<>, std::tuple<>> { - using type = std::tuple<>; +template <> +struct MergeProperties<properties_type_list<>, properties_type_list<>> { + using type = properties_type_list<>; }; template <typename... LHSPropertyTs> -struct MergeProperties<std::tuple<LHSPropertyTs...>, std::tuple<>> { - using type = std::tuple<LHSPropertyTs...>; +struct MergeProperties<properties_type_list<LHSPropertyTs...>, + properties_type_list<>> { + using type = properties_type_list<LHSPropertyTs...>; }; template <typename... RHSPropertyTs> -struct MergeProperties<std::tuple<>, std::tuple<RHSPropertyTs...>> { - using type = std::tuple<RHSPropertyTs...>; +struct MergeProperties<properties_type_list<>, + properties_type_list<RHSPropertyTs...>> { + using type = properties_type_list<RHSPropertyTs...>; }; // Identical properties are allowed, but only one will carry over. template <typename PropertyT, typename... LHSPropertyTs, typename... RHSPropertyTs> -struct MergeProperties<std::tuple<PropertyT, LHSPropertyTs...>, - std::tuple<PropertyT, RHSPropertyTs...>> { +struct MergeProperties<properties_type_list<PropertyT, LHSPropertyTs...>, + properties_type_list<PropertyT, RHSPropertyTs...>> { using merge_tails = - typename MergeProperties<std::tuple<LHSPropertyTs...>, - std::tuple<RHSPropertyTs...>>::type; - using type = typename PrependTuple<PropertyT, merge_tails>::type; + typename MergeProperties<properties_type_list<LHSPropertyTs...>, + properties_type_list<RHSPropertyTs...>>::type; + using type = typename PrependProperty<PropertyT, merge_tails>::type; }; template <typename... LHSPropertyTs, typename... RHSPropertyTs> -struct MergeProperties<std::tuple<LHSPropertyTs...>, - std::tuple<RHSPropertyTs...>> { - using l_head = GetFirstType<LHSPropertyTs...>; - using r_head = GetFirstType<RHSPropertyTs...>; +struct MergeProperties<properties_type_list<LHSPropertyTs...>, + properties_type_list<RHSPropertyTs...>> { + using l_head = nth_type_t<0, LHSPropertyTs...>; + using r_head = nth_type_t<0, RHSPropertyTs...>; static_assert( PropertyID<l_head>::value != PropertyID<r_head>::value, "Failed to merge property lists due to conflicting properties."); static constexpr bool left_has_min = PropertyID<l_head>::value < PropertyID<r_head>::value; - using l_split = HeadSplit<std::tuple<LHSPropertyTs...>, left_has_min>; - using r_split = HeadSplit<std::tuple<RHSPropertyTs...>, !left_has_min>; + using l_split = + HeadSplit<properties_type_list<LHSPropertyTs...>, left_has_min>; + using r_split = + HeadSplit<properties_type_list<RHSPropertyTs...>, !left_has_min>; using min = typename SelectNonVoid<typename l_split::htype, typename r_split::htype>::type; using merge_tails = typename MergeProperties<typename l_split::ttype, typename r_split::ttype>::type; - using type = typename PrependTuple<min, merge_tails>::type; + using type = typename PrependProperty<min, merge_tails>::type; }; //****************************************************************************** diff --git a/sycl/include/sycl/ext/oneapi/properties/property_value.hpp b/sycl/include/sycl/ext/oneapi/properties/property_value.hpp index 629ae794a0b1c..024819d3f77c1 100644 --- a/sycl/include/sycl/ext/oneapi/properties/property_value.hpp +++ b/sycl/include/sycl/ext/oneapi/properties/property_value.hpp @@ -79,6 +79,9 @@ template <typename V, typename O> struct is_property_value_of<V, O, std::void_t<typename V::key_t>> : is_property_key_of<typename V::key_t, O> {}; +template <typename V> +inline constexpr bool is_property_value_v = is_property_value<V>::value; + namespace detail { // Specialization of PropertyID for propagating IDs through property_value. diff --git a/sycl/test/check_device_code/group_load.cpp b/sycl/test/check_device_code/group_load.cpp index ef389cb02da45..c7f159f0ca81d 100644 --- a/sycl/test/check_device_code/group_load.cpp +++ b/sycl/test/check_device_code/group_load.cpp @@ -41,15 +41,15 @@ using plain_global_ptr = typename sycl::detail::DecoratedType< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, naive_blocked>( sycl::sub_group, plain_global_ptr<int>, int &, naive_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_SL_RSM_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] comdat !srcloc [[META5:![0-9]+]] !sycl_fixed_targets [[META6:![0-9]+]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_SL_RSM_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] comdat !srcloc [[META6:![0-9]+]] !sycl_fixed_targets [[META7:![0-9]+]] { // CHECK-NEXT: entry: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4:[0-9]+]] // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7:![0-9]+]] -// CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(4) [[OUT]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8:![0-9]+]] +// CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(4) [[OUT]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -57,8 +57,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, opt_blocked>( sycl::sub_group, plain_global_ptr<int>, int &, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_RSN_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_RSO_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I]]) @@ -70,8 +70,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, full_group_blocked>( sycl::sub_group, plain_global_ptr<int>, int &, full_group_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESM_SK_RSL_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.2") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_SL_RSM_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.12") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I]]) @@ -87,19 +87,19 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load<sycl::sub_group, accessor_iter_t, int, full_group_blocked>( sycl::sub_group, accessor_iter_t, int &, full_group_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_RSN_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.2") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSD_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_RSO_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.12") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[AGG_TMP1_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA11:![0-9]+]] +// CHECK-NEXT: [[AGG_TMP1_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA12:![0-9]+]] // CHECK-NEXT: [[AGG_TMP1_SROA_2_0_IN_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[IN_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP1_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP1_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13:![0-9]+]] +// CHECK-NEXT: [[AGG_TMP1_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP1_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14:![0-9]+]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[CONV3_I_I:%.*]] = sext i32 [[TMP0]] to i64 // CHECK-NEXT: [[TMP1:%.*]] = getelementptr i32, ptr addrspace(4) [[AGG_TMP1_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP1_SROA_2_0_COPYLOAD]] // CHECK-NEXT: [[ADD_PTR_I_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[TMP1]], i64 [[CONV3_I_I]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA7]] -// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(4) [[OUT]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA8]] +// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(4) [[OUT]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -108,12 +108,12 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, accessor_iter_t, int, opt_blocked>(sycl::sub_group, accessor_iter_t, int &, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_SO_RSP_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSD_INS3_21contiguous_memory_keyEJEEENSD_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_RSQ_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[AGG_TMP1_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[AGG_TMP1_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[AGG_TMP1_SROA_2_0_IN_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[IN_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP1_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP1_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13]] +// CHECK-NEXT: [[AGG_TMP1_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP1_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14]] // CHECK-NEXT: [[ADD_PTR_I_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[AGG_TMP1_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP1_SROA_2_0_COPYLOAD]] // CHECK-NEXT: [[CMP_I_I_I:%.*]] = icmp ne ptr addrspace(4) [[ADD_PTR_I_I_I]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I_I]]) @@ -125,15 +125,15 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[CONV3_I_I:%.*]] = sext i32 [[TMP0]] to i64 // CHECK-NEXT: [[ADD_PTR_I_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[ADD_PTR_I_I_I]], i64 [[CONV3_I_I]] -// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA7]] -// CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(4) [[OUT]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA8]] +// CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(4) [[OUT]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM1ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_SO_NS0_4SPANISP_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM1ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSD_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSD_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: if.end.i: // CHECK-NEXT: [[CALL6_I:%.*]] = tail call spir_func noundef i32 @_Z30__spirv_SubgroupBlockReadINTELIjET_PU3AS1Kj(ptr addrspace(1) noundef nonnull [[CALL_I_I_I_I]]) #[[ATTR4]] // CHECK-NEXT: store i32 [[CALL6_I]], ptr addrspace(4) [[OUT]], align 4 -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM1ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_SO_NS0_4SPANISP_XT2_EEET3__EXIT]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm1ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_SO_NS0_4spanISP_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM1ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSD_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSD_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm1ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSD_INS3_21contiguous_memory_keyEJEEENSD_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: ret void // Run-time alignment check is needed if type's alignment is less than BlockRead @@ -141,8 +141,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<char>, char, opt_blocked>( sycl::sub_group, plain_global_ptr<char>, char &, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_RSN_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 1 dereferenceable(1) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_RSO_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr addrspace(4) noundef align 1 dereferenceable(1) [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I]]) @@ -155,15 +155,15 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< // CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[IDXPROM_I_I:%.*]] = sext i32 [[TMP1]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds i8, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I_I]] -// CHECK-NEXT: [[TMP2:%.*]] = load i8, ptr addrspace(1) [[ARRAYIDX_I_I]], align 1, !tbaa [[TBAA15:![0-9]+]] -// CHECK-NEXT: store i8 [[TMP2]], ptr addrspace(4) [[OUT]], align 1, !tbaa [[TBAA15]] +// CHECK-NEXT: [[TMP2:%.*]] = load i8, ptr addrspace(1) [[ARRAYIDX_I_I]], align 1, !tbaa [[TBAA16:![0-9]+]] +// CHECK-NEXT: store i8 [[TMP2]], ptr addrspace(4) [[OUT]], align 1, !tbaa [[TBAA16]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1CCLM1ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESO_SM_NS0_4SPANISN_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1CCLM1ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESP_SN_NS0_4SPANISO_XT2_EEET3__EXIT:%.*]] // CHECK: if.end.i: // CHECK-NEXT: [[CALL4_I:%.*]] = tail call spir_func noundef zeroext i8 @_Z30__spirv_SubgroupBlockReadINTELIhET_PU3AS1Kh(ptr addrspace(1) noundef nonnull [[IN_PTR]]) #[[ATTR4]] // CHECK-NEXT: store i8 [[CALL4_I]], ptr addrspace(4) [[OUT]], align 1 -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1CCLM1ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESO_SM_NS0_4SPANISN_XT2_EEET3__EXIT]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccLm1ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1CCLM1ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESP_SN_NS0_4SPANISO_XT2_EEET3__EXIT]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccLm1ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_.exit: // CHECK-NEXT: ret void // Four shorts in blocked data layout could be loaded as a single 64-bit @@ -171,8 +171,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<short>, short, 4, opt_blocked>( sycl::sub_group, plain_global_ptr<short>, span<short, 4>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.5") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16:![0-9]+]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.15") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17:![0-9]+]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I]]) @@ -181,32 +181,32 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META17:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META18:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = shl i32 [[TMP3]], 2 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I14:%.*]] = icmp samesign ult i32 [[I_0_I]], 4 -// CHECK-NEXT: br i1 [[CMP_I14]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1SSLM4ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I14]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1SSLM4ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ADD_I_I:%.*]] = or disjoint i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i16, ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA20:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i16, ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA21:![0-9]+]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: store i16 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA20]] +// CHECK-NEXT: store i16 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA21]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP22:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP23:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: [[CALL4:%.*]] = tail call spir_func noundef i64 @_Z30__spirv_SubgroupBlockReadINTELImET_PU3AS1Km(ptr addrspace(1) noundef nonnull [[IN_PTR]]) #[[ATTR4]] -// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA24:![0-9]+]] +// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA25:![0-9]+]] // CHECK-NEXT: store i64 [[CALL4]], ptr addrspace(4) [[TMP5]], align 2 // CHECK-NEXT: br label [[CLEANUP]] // CHECK: cleanup: @@ -216,30 +216,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 3, opt_blocked>( sycl::sub_group, plain_global_ptr<int>, span<int, 3>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.6") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.16") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META26:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META27:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP2]], 3 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 3 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM3ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM3ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP29:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP30:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -248,30 +248,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 4, opt_blocked>( sycl::sub_group, plain_global_ptr<int>, span<int, 4>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm4ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.7") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm4ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.17") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META30:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META31:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = shl i32 [[TMP2]], 2 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 4 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM4ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM4ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ADD_I_I:%.*]] = or disjoint i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP33:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm4ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP34:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm4ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -279,30 +279,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 7, opt_blocked>( sycl::sub_group, plain_global_ptr<int>, span<int, 7>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm7ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.8") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm7ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.18") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META34:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META35:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP2]], 7 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 7 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM7ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM7ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP37:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm7ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP38:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm7ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -313,13 +313,13 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 2, naive_striped>( sycl::sub_group, plain_global_ptr<int>, span<int, 2>, naive_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_SL_NS0_4spanISM_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.10") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_SL_NS0_4spanISM_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.20") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META38:![0-9]+]] -// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META41:![0-9]+]] -// CHECK-NEXT: [[TMP2:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA44:![0-9]+]] +// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META39:![0-9]+]] +// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META42:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA45:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] @@ -334,23 +334,23 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< // CHECK-NEXT: [[ADD_I:%.*]] = add i32 [[TMP0]], [[MUL_I]] // CHECK-NEXT: [[IDXPROM:%.*]] = sext i32 [[ADD_I]] to i64 // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP2]], i64 [[CONV]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(4) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP46:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP47:![0-9]+]] // Check that optimized implementation is selected. template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 2, opt_striped>( sycl::sub_group, plain_global_ptr<int>, span<int, 2>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: cleanup: // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I]]) // CHECK-NEXT: [[CALL4:%.*]] = tail call spir_func noundef <2 x i32> @_Z30__spirv_SubgroupBlockReadINTELIDv2_jET_PU3AS1Kj(ptr addrspace(1) noundef nonnull [[IN_PTR]]) #[[ATTR4]] -// CHECK-NEXT: [[TMP0:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA44]] +// CHECK-NEXT: [[TMP0:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA45]] // CHECK-NEXT: store <2 x i32> [[CALL4]], ptr addrspace(4) [[TMP0]], align 4 // CHECK-NEXT: ret void @@ -358,13 +358,13 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 2, full_group_striped>( sycl::sub_group, plain_global_ptr<int>, span<int, 2>, full_group_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESM_SK_NS0_4spanISL_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.13") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_SL_NS0_4spanISM_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.27") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: cleanup: // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I]]) // CHECK-NEXT: [[CALL4:%.*]] = tail call spir_func noundef <2 x i32> @_Z30__spirv_SubgroupBlockReadINTELIDv2_jET_PU3AS1Kj(ptr addrspace(1) noundef nonnull [[IN_PTR]]) #[[ATTR4]] -// CHECK-NEXT: [[TMP0:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA44]] +// CHECK-NEXT: [[TMP0:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA45]] // CHECK-NEXT: store <2 x i32> [[CALL4]], ptr addrspace(4) [[TMP0]], align 4 // CHECK-NEXT: ret void @@ -375,35 +375,35 @@ using accessor_iter_t = accessor<int, 1, access_mode::read, target::device, template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, accessor_iter_t, int, 2, full_group_striped>( sycl::sub_group, accessor_iter_t, span<int, 2>, full_group_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.13") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSD_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.27") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[AGG_TMP3_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[AGG_TMP3_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[AGG_TMP3_SROA_2_0_IN_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[IN_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP3_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP3_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13]] -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[AGG_TMP3_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP3_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META47:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META50:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META48:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META51:![0-9]+]] // CHECK-NEXT: [[TMP4:%.*]] = getelementptr i32, ptr addrspace(4) [[AGG_TMP3_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP3_SROA_2_0_COPYLOAD]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 2 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM2ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSC_INS3_14FULL_GROUP_KEYEJEEENSC_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM2ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSD_INS3_14FULL_GROUP_KEYEJEEENSD_INSB_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[MUL_I_I:%.*]] = mul nuw nsw i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[CONV3_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ADD_PTR_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[TMP4]], i64 [[CONV3_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP53:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_14full_group_keyEJEEENSC_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP54:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSD_INS3_14full_group_keyEJEEENSD_INSB_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -413,12 +413,12 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< accessor_iter_t, span<int, 2>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_SO_NS0_4spanISP_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSD_INS3_21contiguous_memory_keyEJEEENSD_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[AGG_TMP_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[AGG_TMP_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[IN_PTR]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[AGG_TMP_SROA_2_0_IN_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[IN_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13]] +// CHECK-NEXT: [[AGG_TMP_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP_SROA_2_0_IN_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14]] // CHECK-NEXT: [[ADD_PTR_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[AGG_TMP_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP_SROA_2_0_COPYLOAD]] // CHECK-NEXT: [[CMP_I_I:%.*]] = icmp ne ptr addrspace(4) [[ADD_PTR_I_I]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I]]) @@ -426,33 +426,33 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< // CHECK-NEXT: [[TOBOOL_NOT:%.*]] = icmp eq ptr addrspace(1) [[CALL_I_I_I]], null // CHECK-NEXT: br i1 [[TOBOOL_NOT]], label [[IF_THEN:%.*]], label [[IF_END:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META54:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META57:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META55:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META58:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 2 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM2ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEENSC_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEEST_SR_NS0_4SPANISS_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPENS0_6DETAIL17ACCESSOR_ITERATORIKILI1EEEILM2ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSD_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSD_INS3_14FULL_GROUP_KEYEJEEENSD_INSB_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEEST_SR_NS0_4SPANISS_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[MUL_I_I:%.*]] = mul nuw nsw i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[CONV3_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ADD_PTR_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[ADD_PTR_I_I]], i64 [[CONV3_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP60:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEENSC_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeEST_SR_NS0_4spanISS_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP61:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupENS0_6detail17accessor_iteratorIKiLi1EEEiLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSD_INS3_21contiguous_memory_keyEJEEENSD_INS3_14full_group_keyEJEEENSD_INSB_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeEST_SR_NS0_4spanISS_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: [[CALL6:%.*]] = tail call spir_func noundef <2 x i32> @_Z30__spirv_SubgroupBlockReadINTELIDv2_jET_PU3AS1Kj(ptr addrspace(1) noundef nonnull [[CALL_I_I_I]]) #[[ATTR4]] -// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA44]] +// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA45]] // CHECK-NEXT: store <2 x i32> [[CALL6]], ptr addrspace(4) [[TMP5]], align 4 // CHECK-NEXT: br label [[CLEANUP]] // CHECK: cleanup: @@ -463,8 +463,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<char>, char, 2, opt_striped>( sycl::sub_group, plain_global_ptr<char>, span<char, 2>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.15") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.29") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I]]) @@ -473,33 +473,33 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META61:![0-9]+]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META64:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META62:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META65:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I14:%.*]] = icmp samesign ult i32 [[I_0_I]], 2 -// CHECK-NEXT: br i1 [[CMP_I14]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1CCLM2ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I14]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1CCLM2ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[MUL_I_I:%.*]] = mul nuw nsw i32 [[TMP4]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP3]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i8, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i8, ptr addrspace(1) [[ARRAYIDX_I]], align 1, !tbaa [[TBAA15]] +// CHECK-NEXT: [[TMP5:%.*]] = load i8, ptr addrspace(1) [[ARRAYIDX_I]], align 1, !tbaa [[TBAA16]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i8, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: store i8 [[TMP5]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 1, !tbaa [[TBAA15]] +// CHECK-NEXT: store i8 [[TMP5]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 1, !tbaa [[TBAA16]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP67:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccLm2ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP68:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ccLm2ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: [[CALL4:%.*]] = tail call spir_func noundef <2 x i8> @_Z30__spirv_SubgroupBlockReadINTELIDv2_hET_PU3AS1Kh(ptr addrspace(1) noundef nonnull [[IN_PTR]]) #[[ATTR4]] -// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA68:![0-9]+]] +// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA69:![0-9]+]] // CHECK-NEXT: store <2 x i8> [[CALL4]], ptr addrspace(4) [[TMP6]], align 1 // CHECK-NEXT: br label [[CLEANUP]] // CHECK: cleanup: @@ -510,8 +510,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<short>, short, 4, opt_striped>( sycl::sub_group, plain_global_ptr<short>, span<short, 4>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.5") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.15") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[IN_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I]]) @@ -520,33 +520,33 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META70:![0-9]+]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META73:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META71:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META74:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I14:%.*]] = icmp samesign ult i32 [[I_0_I]], 4 -// CHECK-NEXT: br i1 [[CMP_I14]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1SSLM4ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I14]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1SSLM4ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP4]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP3]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i16, ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA20]] +// CHECK-NEXT: [[TMP5:%.*]] = load i16, ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA21]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: store i16 [[TMP5]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA20]] +// CHECK-NEXT: store i16 [[TMP5]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA21]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP76:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP77:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1ssLm4ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: [[CALL4:%.*]] = tail call spir_func noundef <4 x i16> @_Z30__spirv_SubgroupBlockReadINTELIDv4_tET_PU3AS1Kt(ptr addrspace(1) noundef nonnull [[IN_PTR]]) #[[ATTR4]] -// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA24]] +// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[OUT]], align 8, !tbaa [[TBAA25]] // CHECK-NEXT: store <4 x i16> [[CALL4]], ptr addrspace(4) [[TMP6]], align 2 // CHECK-NEXT: br label [[CLEANUP]] // CHECK: cleanup: @@ -556,31 +556,31 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 3, opt_striped>( sycl::sub_group, plain_global_ptr<int>, span<int, 3>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.6") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.16") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META77:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META80:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META78:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META81:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 3 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM3ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM3ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP83:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP84:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm3ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -588,31 +588,31 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 16, opt_striped>( sycl::sub_group, plain_global_ptr<int>, span<int, 16>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm16ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.16") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm16ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.30") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META84:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META87:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META85:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META88:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 16 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM16ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM16ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP90:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm16ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP91:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm16ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void @@ -620,30 +620,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_load< sycl::sub_group, plain_global_ptr<int>, int, 11, opt_striped>( sycl::sub_group, plain_global_ptr<int>, span<int, 11>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm11ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_SM_NS0_4spanISN_XT2_EEET3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.17") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm11ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_SN_NS0_4spanISO_XT2_EEET3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(1) noundef [[IN_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::span.31") align 8 [[OUT:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META17]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[OUT]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META91:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META94:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META92:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META95:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 11 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM11ENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL10GROUP_LOADINS0_9SUB_GROUPEPU3AS1IILM11ENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE17VERIFY_LOAD_TYPESIT0_T1_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_SP_NS0_4SPANISQ_XT2_EEET3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[IN_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP97:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm11ENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP98:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental10group_loadINS0_9sub_groupEPU3AS1iiLm11ENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE17verify_load_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_SP_NS0_4spanISQ_XT2_EEET3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR4]] // CHECK-NEXT: ret void diff --git a/sycl/test/check_device_code/group_store.cpp b/sycl/test/check_device_code/group_store.cpp index 60c241f91ebeb..7d7fa06ad0763 100644 --- a/sycl/test/check_device_code/group_store.cpp +++ b/sycl/test/check_device_code/group_store.cpp @@ -41,15 +41,15 @@ using plain_global_ptr = typename sycl::detail::DecoratedType< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, plain_global_ptr<int>, naive_blocked>( sycl::sub_group, const int &, plain_global_ptr<int>, naive_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_RKSL_SM_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] comdat !srcloc [[META5:![0-9]+]] !sycl_fixed_targets [[META6:![0-9]+]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_RKSL_SM_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] comdat !srcloc [[META6:![0-9]+]] !sycl_fixed_targets [[META7:![0-9]+]] { // CHECK-NEXT: entry: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5:[0-9]+]] // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7:![0-9]+]] -// CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8:![0-9]+]] +// CHECK-NEXT: store i32 [[TMP1]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -57,8 +57,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, plain_global_ptr<int>, opt_blocked>( sycl::sub_group, const int &, plain_global_ptr<int>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_RKSM_SN_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_RKSN_SO_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I_I:%.*]] = icmp ne ptr addrspace(1) [[OUT_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I]]) @@ -71,23 +71,23 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[IDXPROM_I_I:%.*]] = sext i32 [[TMP1]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I_I]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7]] -// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(1) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8]] +// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(1) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESP_NS0_4SPANISN_XT1_EEESO_T3__EXIT:%.*]] +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_NS0_4SPANISO_XT1_EEESP_T3__EXIT:%.*]] // CHECK: if.end.i: -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELIjEvPU3AS1jT_(ptr addrspace(1) noundef nonnull [[OUT_PTR]], i32 noundef [[TMP3]]) #[[ATTR5]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESP_NS0_4SPANISN_XT1_EEESO_T3__EXIT]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKiLm1EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_.exit: +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_NS0_4SPANISO_XT1_EEESP_T3__EXIT]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKiLm1EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_NS0_4spanISO_XT1_EEESP_T3_.exit: // CHECK-NEXT: ret void // Check that contiguous_memory can be auto-detected. template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, plain_global_ptr<int>, full_group_blocked>( sycl::sub_group, const int &, plain_global_ptr<int>, full_group_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESM_RKSK_SL_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.2") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_RKSL_SM_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.12") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[CMP_I_I:%.*]] = icmp ne ptr addrspace(1) [[OUT_PTR]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I]]) @@ -100,15 +100,15 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[IDXPROM_I_I:%.*]] = sext i32 [[TMP1]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I_I]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7]] -// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(1) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8]] +// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(1) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESN_NS0_4SPANISL_XT1_EEESM_T3__EXIT:%.*]] +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESO_NS0_4SPANISM_XT1_EEESN_T3__EXIT:%.*]] // CHECK: if.end.i: -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELIjEvPU3AS1jT_(ptr addrspace(1) noundef nonnull [[OUT_PTR]], i32 noundef [[TMP3]]) #[[ATTR5]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESN_NS0_4SPANISL_XT1_EEESM_T3__EXIT]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKiLm1EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_NS0_4spanISL_XT1_EEESM_T3_.exit: +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESO_NS0_4SPANISM_XT1_EEESN_T3__EXIT]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKiLm1EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_.exit: // CHECK-NEXT: ret void // SYCL 2020's accessor can't be statically known to be contiguous. @@ -118,19 +118,19 @@ using accessor_iter_t = accessor<int, 1, access_mode::write, target::device, template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, accessor_iter_t, full_group_blocked>( sycl::sub_group, const int &, accessor_iter_t, full_group_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiNS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_RKSL_SM_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.2") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiNS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_RKSM_SN_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.12") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[AGG_TMP2_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA11:![0-9]+]] +// CHECK-NEXT: [[AGG_TMP2_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA12:![0-9]+]] // CHECK-NEXT: [[AGG_TMP2_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[OUT_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP2_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP2_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13:![0-9]+]] +// CHECK-NEXT: [[AGG_TMP2_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP2_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14:![0-9]+]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[CONV5_I_I:%.*]] = sext i32 [[TMP0]] to i64 // CHECK-NEXT: [[TMP1:%.*]] = getelementptr i32, ptr addrspace(4) [[AGG_TMP2_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP2_SROA_2_0_COPYLOAD]] // CHECK-NEXT: [[ADD_PTR_I_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[TMP1]], i64 [[CONV5_I_I]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7]] -// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8]] +// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -140,20 +140,20 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< const int &, accessor_iter_t, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiNS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_RKSN_SO_T2_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META5]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiNS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T1_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_RKSO_SP_T2_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr addrspace(4) noundef align 4 dereferenceable(4) [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPERTIES:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META6]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[AGG_TMP2_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[AGG_TMP2_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[AGG_TMP2_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[OUT_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP2_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP2_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13]] +// CHECK-NEXT: [[AGG_TMP2_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP2_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14]] // CHECK-NEXT: [[ADD_PTR_I_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[AGG_TMP2_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP2_SROA_2_0_COPYLOAD]] // CHECK-NEXT: [[CMP_I_I_I:%.*]] = icmp ne ptr addrspace(4) [[ADD_PTR_I_I_I]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I_I]]) // CHECK-NEXT: [[TMP0:%.*]] = ptrtoint ptr addrspace(4) [[ADD_PTR_I_I_I]] to i64 // CHECK-NEXT: [[REM_I_I_I:%.*]] = and i64 [[TMP0]], 15 // CHECK-NEXT: [[CMP1_I_I_I:%.*]] = icmp eq i64 [[REM_I_I_I]], 0 -// CHECK-NEXT: br i1 [[CMP1_I_I_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL6DETAIL16GET_BLOCK_OP_PTRILI16ELM1ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEEDAT1_T2__EXIT_I:%.*]], label [[IF_THEN_I:%.*]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental6detail16get_block_op_ptrILi16ELm1ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEDaT1_T2_.exit.i: +// CHECK-NEXT: br i1 [[CMP1_I_I_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL6DETAIL16GET_BLOCK_OP_PTRILI16ELM1ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESINS4_20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEEEDAT1_T2__EXIT_I:%.*]], label [[IF_THEN_I:%.*]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental6detail16get_block_op_ptrILi16ELm1ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS4_20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEEDaT1_T2_.exit.i: // CHECK-NEXT: [[CALL_I_I_I_I:%.*]] = tail call spir_func noundef ptr addrspace(1) @_Z41__spirv_GenericCastToPtrExplicit_ToGlobalPvi(ptr addrspace(4) noundef nonnull [[ADD_PTR_I_I_I]], i32 noundef 5) #[[ATTR6:[0-9]+]] // CHECK-NEXT: [[TOBOOL_NOT_I:%.*]] = icmp eq ptr addrspace(1) [[CALL_I_I_I_I]], null // CHECK-NEXT: br i1 [[TOBOOL_NOT_I]], label [[IF_THEN_I]], label [[IF_END_I:%.*]] @@ -162,15 +162,15 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4 // CHECK-NEXT: [[CONV5_I_I:%.*]] = sext i32 [[TMP1]] to i64 // CHECK-NEXT: [[ADD_PTR_I_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[ADD_PTR_I_I_I]], i64 [[CONV5_I_I]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7]] -// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8]] +// CHECK-NEXT: store i32 [[TMP2]], ptr addrspace(4) [[ADD_PTR_I_I_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_NS0_4SPANISO_XT1_EEESP_T3__EXIT:%.*]] +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSD_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSD_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: if.end.i: -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[IN]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELIjEvPU3AS1jT_(ptr addrspace(1) noundef nonnull [[CALL_I_I_I_I]], i32 noundef [[TMP3]]) #[[ATTR5]] -// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_NS0_4SPANISO_XT1_EEESP_T3__EXIT]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKiLm1ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_NS0_4spanISO_XT1_EEESP_T3_.exit: +// CHECK-NEXT: br label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKILM1ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSD_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSD_INS3_14FULL_GROUP_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKiLm1ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSD_INS3_21contiguous_memory_keyEJEEENSD_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: ret void // Four shorts in blocked data layout could be stored as a single 64-bit @@ -178,8 +178,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, short, 4, plain_global_ptr<short>, opt_blocked>( sycl::sub_group, span<short, 4>, plain_global_ptr<short>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.4") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15:![0-9]+]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.14") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16:![0-9]+]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[VALUES:%.*]] = alloca [4 x i16], align 2 // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[OUT_PTR]], null @@ -189,50 +189,50 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META16:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META17:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = shl i32 [[TMP3]], 2 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I19:%.*]] = icmp samesign ult i32 [[I_0_I]], 4 -// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPESLM4EPU3AS1SNS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPESLM4EPU3AS1SNS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA19:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA20:![0-9]+]] // CHECK-NEXT: [[ADD_I_I:%.*]] = or disjoint i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i16 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: store i16 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP21:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP22:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7:[0-9]+]] -// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA23:![0-9]+]] +// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA24:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[IF_END]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] // CHECK-NEXT: [[CMP:%.*]] = icmp samesign ult i32 [[I_0]], 4 // CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_COND_CLEANUP:%.*]] // CHECK: for.cond.cleanup: -// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr [[VALUES]], align 2, !tbaa [[TBAA25:![0-9]+]] +// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr [[VALUES]], align 2, !tbaa [[TBAA26:![0-9]+]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELImEvPU3AS1mT_(ptr addrspace(1) noundef nonnull [[OUT_PTR]], i64 noundef [[TMP6]]) #[[ATTR5]] // CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] // CHECK-NEXT: br label [[CLEANUP]] // CHECK: for.body: // CHECK-NEXT: [[CONV:%.*]] = zext nneg i32 [[I_0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I20:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP5]], i64 [[CONV]] -// CHECK-NEXT: [[TMP7:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I20]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: [[TMP7:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I20]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x i16], ptr [[VALUES]], i64 0, i64 [[CONV]] -// CHECK-NEXT: store i16 [[TMP7]], ptr [[ARRAYIDX]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: store i16 [[TMP7]], ptr [[ARRAYIDX]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP26:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP27:![0-9]+]] // CHECK: cleanup: // CHECK-NEXT: ret void @@ -241,8 +241,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, const short, 4, plain_global_ptr<short>, opt_blocked>( sycl::sub_group, span<const short, 4>, plain_global_ptr<short>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKsLm4EPU3AS1sNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.5") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKsLm4EPU3AS1sNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_NS0_4spanISO_XT1_EEESP_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.15") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[VALUES:%.*]] = alloca [4 x i16], align 2 // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[OUT_PTR]], null @@ -252,50 +252,50 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META27:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META28:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = shl i32 [[TMP3]], 2 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I19:%.*]] = icmp samesign ult i32 [[I_0_I]], 4 -// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKSLM4EPU3AS1SNS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESS_NS0_4SPANISQ_XT1_EEESR_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEKSLM4EPU3AS1SNS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEENSC_INSA_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESS_NS0_4SPANISQ_XT1_EEESR_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: [[TMP4:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[ADD_I_I:%.*]] = or disjoint i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i16 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: store i16 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP30:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKsLm4EPU3AS1sNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESS_NS0_4spanISQ_XT1_EEESR_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP31:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEKsLm4EPU3AS1sNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEENSC_INSA_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESS_NS0_4spanISQ_XT1_EEESR_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] -// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA31:![0-9]+]] +// CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA32:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[IF_END]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] // CHECK-NEXT: [[CMP:%.*]] = icmp samesign ult i32 [[I_0]], 4 // CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_COND_CLEANUP:%.*]] // CHECK: for.cond.cleanup: -// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr [[VALUES]], align 2, !tbaa [[TBAA25]] +// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr [[VALUES]], align 2, !tbaa [[TBAA26]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELImEvPU3AS1mT_(ptr addrspace(1) noundef nonnull [[OUT_PTR]], i64 noundef [[TMP6]]) #[[ATTR5]] // CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] // CHECK-NEXT: br label [[CLEANUP]] // CHECK: for.body: // CHECK-NEXT: [[CONV:%.*]] = zext nneg i32 [[I_0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I20:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP5]], i64 [[CONV]] -// CHECK-NEXT: [[TMP7:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I20]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: [[TMP7:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I20]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x i16], ptr [[VALUES]], i64 0, i64 [[CONV]] -// CHECK-NEXT: store i16 [[TMP7]], ptr [[ARRAYIDX]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: store i16 [[TMP7]], ptr [[ARRAYIDX]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP33:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP34:![0-9]+]] // CHECK: cleanup: // CHECK-NEXT: ret void @@ -303,30 +303,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 3, plain_global_ptr<int>, opt_blocked>( sycl::sub_group, span<int, 3>, plain_global_ptr<int>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.6") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.16") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META34:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META35:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP2]], 3 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 3 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM3EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM3EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP37:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP38:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -335,30 +335,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 4, plain_global_ptr<int>, opt_blocked>( sycl::sub_group, span<int, 4>, plain_global_ptr<int>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm4EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.7") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm4EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.17") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META38:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META39:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = shl i32 [[TMP2]], 2 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 4 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM4EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM4EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ADD_I_I:%.*]] = or disjoint i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP41:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm4EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP42:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm4EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -366,30 +366,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 7, plain_global_ptr<int>, opt_blocked>( sycl::sub_group, span<int, 7>, plain_global_ptr<int>, opt_blocked); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm7EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.8") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.0") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm7EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.18") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.4") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META42:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META43:![0-9]+]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP2]], 7 // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 7 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM7EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM7EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI0EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[MUL_I_I]], [[I_0_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP45:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm7EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP46:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm7EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi0EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -400,13 +400,13 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 2, plain_global_ptr<int>, naive_striped>( sycl::sub_group, span<int, 2>, plain_global_ptr<int>, naive_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_NS0_4spanISL_XT1_EEESM_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.10") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_NS0_4spanISL_XT1_EEESM_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.20") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP0:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA46:![0-9]+]] -// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META48:![0-9]+]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META51:![0-9]+]] +// CHECK-NEXT: [[TMP0:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA47:![0-9]+]] +// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META49:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META52:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] @@ -418,21 +418,21 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK: for.body: // CHECK-NEXT: [[CONV:%.*]] = zext nneg i32 [[I_0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP0]], i64 [[CONV]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I:%.*]] = mul nuw nsw i32 [[TMP2]], [[I_0]] // CHECK-NEXT: [[ADD_I:%.*]] = add i32 [[TMP1]], [[MUL_I]] // CHECK-NEXT: [[IDXPROM:%.*]] = sext i32 [[ADD_I]] to i64 // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM]] -// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP3]], ptr addrspace(1) [[ARRAYIDX]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP54:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP55:![0-9]+]] // Check that optimized implementation is selected. template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 2, plain_global_ptr<int>, opt_striped>( sycl::sub_group, span<int, 2>, plain_global_ptr<int>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[VALUES:%.*]] = alloca [2 x i32], align 4 // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[OUT_PTR]], null @@ -442,51 +442,51 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META55:![0-9]+]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META58:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META56:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META59:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I19:%.*]] = icmp samesign ult i32 [[I_0_I]], 2 -// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul nuw nsw i32 [[TMP4]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP3]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP61:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP62:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] -// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA46]] +// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA47]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[IF_END]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] // CHECK-NEXT: [[CMP:%.*]] = icmp samesign ult i32 [[I_0]], 2 // CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_COND_CLEANUP:%.*]] // CHECK: for.cond.cleanup: -// CHECK-NEXT: [[TMP7:%.*]] = load <2 x i32>, ptr [[VALUES]], align 4, !tbaa [[TBAA25]] +// CHECK-NEXT: [[TMP7:%.*]] = load <2 x i32>, ptr [[VALUES]], align 4, !tbaa [[TBAA26]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELIDv2_jEvPU3AS1jT_(ptr addrspace(1) noundef nonnull [[OUT_PTR]], <2 x i32> noundef [[TMP7]]) #[[ATTR5]] // CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] // CHECK-NEXT: br label [[CLEANUP]] // CHECK: for.body: // CHECK-NEXT: [[CONV:%.*]] = zext nneg i32 [[I_0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I20:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP6]], i64 [[CONV]] -// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I20]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I20]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[VALUES]], i64 0, i64 [[CONV]] -// CHECK-NEXT: store i32 [[TMP8]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP8]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP62:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP63:![0-9]+]] // CHECK: cleanup: // CHECK-NEXT: ret void @@ -494,8 +494,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 2, plain_global_ptr<int>, full_group_striped>( sycl::sub_group, span<int, 2>, plain_global_ptr<int>, full_group_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESM_NS0_4spanISK_XT1_EEESL_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.13") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_NS0_4spanISL_XT1_EEESM_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.27") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[VALUES:%.*]] = alloca [2 x i32], align 4 // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[OUT_PTR]], null @@ -505,51 +505,51 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META63:![0-9]+]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META66:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META64:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META67:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I19:%.*]] = icmp samesign ult i32 [[I_0_I]], 2 -// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESP_NS0_4SPANISN_XT1_EEESO_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESP_NS0_4SPANISN_XT1_EEESO_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul nuw nsw i32 [[TMP4]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP3]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP69:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP70:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] -// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA46]] +// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA47]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[IF_END]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] // CHECK-NEXT: [[CMP:%.*]] = icmp samesign ult i32 [[I_0]], 2 // CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_COND_CLEANUP:%.*]] // CHECK: for.cond.cleanup: -// CHECK-NEXT: [[TMP7:%.*]] = load <2 x i32>, ptr [[VALUES]], align 4, !tbaa [[TBAA25]] +// CHECK-NEXT: [[TMP7:%.*]] = load <2 x i32>, ptr [[VALUES]], align 4, !tbaa [[TBAA26]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELIDv2_jEvPU3AS1jT_(ptr addrspace(1) noundef nonnull [[OUT_PTR]], <2 x i32> noundef [[TMP7]]) #[[ATTR5]] // CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] // CHECK-NEXT: br label [[CLEANUP]] // CHECK: for.body: // CHECK-NEXT: [[CONV:%.*]] = zext nneg i32 [[I_0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I20:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP6]], i64 [[CONV]] -// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I20]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I20]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[VALUES]], i64 0, i64 [[CONV]] -// CHECK-NEXT: store i32 [[TMP8]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP8]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP70:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP71:![0-9]+]] // CHECK: cleanup: // CHECK-NEXT: ret void @@ -560,35 +560,35 @@ using accessor_iter_t = accessor<int, 1, access_mode::write, target::device, template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 2, accessor_iter_t, full_group_striped>( sycl::sub_group, span<int, 2>, accessor_iter_t, full_group_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESN_NS0_4spanISL_XT1_EEESM_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.13") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.27") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] -// CHECK-NEXT: [[AGG_TMP4_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] +// CHECK-NEXT: [[AGG_TMP4_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[AGG_TMP4_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[OUT_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP4_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP4_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13]] +// CHECK-NEXT: [[AGG_TMP4_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP4_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META71:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META74:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META72:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META75:![0-9]+]] // CHECK-NEXT: [[TMP4:%.*]] = getelementptr i32, ptr addrspace(4) [[AGG_TMP4_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP4_SROA_2_0_COPYLOAD]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 2 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_NS0_4SPANISO_XT1_EEESP_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSC_INS3_14FULL_GROUP_KEYEJEEENSC_INSA_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESQ_NS0_4SPANISO_XT1_EEESP_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul nuw nsw i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[CONV5_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ADD_PTR_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[TMP4]], i64 [[CONV5_I]] -// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP77:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_14full_group_keyEJEEENSB_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_NS0_4spanISO_XT1_EEESP_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP78:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_14full_group_keyEJEEENSC_INSA_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_NS0_4spanISO_XT1_EEESP_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -598,70 +598,70 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< span<int, 2>, accessor_iter_t, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.9") align 8 [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESQ_NS0_4spanISO_XT1_EEESP_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.19") align 8 [[IN:%.*]], ptr noundef byval(%"class.sycl::_V1::detail::accessor_iterator") align 8 [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[VALUES:%.*]] = alloca [2 x i32], align 4 -// CHECK-NEXT: [[AGG_TMP_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[AGG_TMP_SROA_0_0_COPYLOAD:%.*]] = load ptr addrspace(4), ptr [[OUT_PTR]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[AGG_TMP_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX:%.*]] = getelementptr inbounds i8, ptr [[OUT_PTR]], i64 8 -// CHECK-NEXT: [[AGG_TMP_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA13]] +// CHECK-NEXT: [[AGG_TMP_SROA_2_0_COPYLOAD:%.*]] = load i64, ptr [[AGG_TMP_SROA_2_0_OUT_PTR_ASCAST_SROA_IDX]], align 8, !tbaa [[TBAA14]] // CHECK-NEXT: [[ADD_PTR_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[AGG_TMP_SROA_0_0_COPYLOAD]], i64 [[AGG_TMP_SROA_2_0_COPYLOAD]] // CHECK-NEXT: [[CMP_I_I:%.*]] = icmp ne ptr addrspace(4) [[ADD_PTR_I_I]], null // CHECK-NEXT: tail call void @llvm.assume(i1 [[CMP_I_I]]) // CHECK-NEXT: [[TMP0:%.*]] = ptrtoint ptr addrspace(4) [[ADD_PTR_I_I]] to i64 // CHECK-NEXT: [[REM_I_I:%.*]] = and i64 [[TMP0]], 15 // CHECK-NEXT: [[CMP1_I_I:%.*]] = icmp eq i64 [[REM_I_I]], 0 -// CHECK-NEXT: br i1 [[CMP1_I_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL6DETAIL16GET_BLOCK_OP_PTRILI16ELM2ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEEDAT1_T2__EXIT:%.*]], label [[IF_THEN:%.*]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental6detail16get_block_op_ptrILi16ELm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEDaT1_T2_.exit: +// CHECK-NEXT: br i1 [[CMP1_I_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL6DETAIL16GET_BLOCK_OP_PTRILI16ELM2ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESINS4_20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEEEEEEEEEDAT1_T2__EXIT:%.*]], label [[IF_THEN:%.*]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental6detail16get_block_op_ptrILi16ELm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS4_20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEEDaT1_T2_.exit: // CHECK-NEXT: [[CALL_I_I_I:%.*]] = tail call spir_func noundef ptr addrspace(1) @_Z41__spirv_GenericCastToPtrExplicit_ToGlobalPvi(ptr addrspace(4) noundef nonnull [[ADD_PTR_I_I]], i32 noundef 5) #[[ATTR6]] // CHECK-NEXT: [[TOBOOL_NOT:%.*]] = icmp eq ptr addrspace(1) [[CALL_I_I_I]], null // CHECK-NEXT: br i1 [[TOBOOL_NOT]], label [[IF_THEN]], label [[IF_END:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META78:![0-9]+]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META81:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META79:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META82:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 2 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESS_NS0_4SPANISQ_XT1_EEESR_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM2ENS0_6DETAIL17ACCESSOR_ITERATORIILI1EEENS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSC_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSC_INS3_14FULL_GROUP_KEYEJEEENSC_INSA_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESS_NS0_4SPANISQ_XT1_EEESR_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP5:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul nuw nsw i32 [[TMP4]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP3]], [[MUL_I_I]] // CHECK-NEXT: [[CONV5_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ADD_PTR_I_I_I:%.*]] = getelementptr i32, ptr addrspace(4) [[ADD_PTR_I_I]], i64 [[CONV5_I]] -// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP5]], ptr addrspace(4) [[ADD_PTR_I_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP84:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESS_NS0_4spanISQ_XT1_EEESR_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP85:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm2ENS0_6detail17accessor_iteratorIiLi1EEENS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSC_INS3_21contiguous_memory_keyEJEEENSC_INS3_14full_group_keyEJEEENSC_INSA_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESS_NS0_4spanISQ_XT1_EEESR_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] -// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA46]] +// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA47]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[IF_END]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] // CHECK-NEXT: [[CMP:%.*]] = icmp samesign ult i32 [[I_0]], 2 // CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_COND_CLEANUP:%.*]] // CHECK: for.cond.cleanup: -// CHECK-NEXT: [[TMP7:%.*]] = load <2 x i32>, ptr [[VALUES]], align 4, !tbaa [[TBAA25]] +// CHECK-NEXT: [[TMP7:%.*]] = load <2 x i32>, ptr [[VALUES]], align 4, !tbaa [[TBAA26]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELIDv2_jEvPU3AS1jT_(ptr addrspace(1) noundef nonnull [[CALL_I_I_I]], <2 x i32> noundef [[TMP7]]) #[[ATTR5]] // CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] // CHECK-NEXT: br label [[CLEANUP]] // CHECK: for.body: // CHECK-NEXT: [[CONV:%.*]] = zext nneg i32 [[I_0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP6]], i64 [[CONV]] -// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[VALUES]], i64 0, i64 [[CONV]] -// CHECK-NEXT: store i32 [[TMP8]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP8]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP85:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP86:![0-9]+]] // CHECK: cleanup: // CHECK-NEXT: ret void @@ -670,8 +670,8 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, short, 4, plain_global_ptr<short>, opt_striped>( sycl::sub_group, span<short, 4>, plain_global_ptr<short>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.4") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.14") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: // CHECK-NEXT: [[VALUES:%.*]] = alloca [4 x i16], align 2 // CHECK-NEXT: [[CMP_I:%.*]] = icmp ne ptr addrspace(1) [[OUT_PTR]], null @@ -681,51 +681,51 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< // CHECK-NEXT: [[CMP1_I_NOT:%.*]] = icmp eq i64 [[REM_I]], 0 // CHECK-NEXT: br i1 [[CMP1_I_NOT]], label [[IF_END:%.*]], label [[IF_THEN:%.*]] // CHECK: if.then: -// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP2:%.*]] = inttoptr i64 [[TMP1]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META86:![0-9]+]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META89:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META87:![0-9]+]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META90:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[IF_THEN]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I19:%.*]] = icmp samesign ult i32 [[I_0_I]], 4 -// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPESLM4EPU3AS1SNS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I19]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPESLM4EPU3AS1SNS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP2]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP5:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: [[TMP5:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I_I]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP4]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP3]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i16 [[TMP5]], ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: store i16 [[TMP5]], ptr addrspace(1) [[ARRAYIDX_I]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP92:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP93:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEsLm4EPU3AS1sNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: br label [[CLEANUP:%.*]] // CHECK: if.end: // CHECK-NEXT: call void @llvm.lifetime.start.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] -// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA23]] +// CHECK-NEXT: [[TMP6:%.*]] = load ptr addrspace(4), ptr [[IN]], align 8, !tbaa [[TBAA24]] // CHECK-NEXT: br label [[FOR_COND:%.*]] // CHECK: for.cond: // CHECK-NEXT: [[I_0:%.*]] = phi i32 [ 0, [[IF_END]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ] // CHECK-NEXT: [[CMP:%.*]] = icmp samesign ult i32 [[I_0]], 4 // CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_COND_CLEANUP:%.*]] // CHECK: for.cond.cleanup: -// CHECK-NEXT: [[TMP7:%.*]] = load <4 x i16>, ptr [[VALUES]], align 2, !tbaa [[TBAA25]] +// CHECK-NEXT: [[TMP7:%.*]] = load <4 x i16>, ptr [[VALUES]], align 2, !tbaa [[TBAA26]] // CHECK-NEXT: tail call spir_func void @_Z31__spirv_SubgroupBlockWriteINTELIDv4_tEvPU3AS1tT_(ptr addrspace(1) noundef nonnull [[OUT_PTR]], <4 x i16> noundef [[TMP7]]) #[[ATTR5]] // CHECK-NEXT: call void @llvm.lifetime.end.p0(i64 8, ptr nonnull [[VALUES]]) #[[ATTR7]] // CHECK-NEXT: br label [[CLEANUP]] // CHECK: for.body: // CHECK-NEXT: [[CONV:%.*]] = zext nneg i32 [[I_0]] to i64 // CHECK-NEXT: [[ARRAYIDX_I20:%.*]] = getelementptr inbounds nuw i16, ptr addrspace(4) [[TMP6]], i64 [[CONV]] -// CHECK-NEXT: [[TMP8:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I20]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: [[TMP8:%.*]] = load i16, ptr addrspace(4) [[ARRAYIDX_I20]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [4 x i16], ptr [[VALUES]], i64 0, i64 [[CONV]] -// CHECK-NEXT: store i16 [[TMP8]], ptr [[ARRAYIDX]], align 2, !tbaa [[TBAA19]] +// CHECK-NEXT: store i16 [[TMP8]], ptr [[ARRAYIDX]], align 2, !tbaa [[TBAA20]] // CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I_0]], 1 -// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP93:![0-9]+]] +// CHECK-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP94:![0-9]+]] // CHECK: cleanup: // CHECK-NEXT: ret void @@ -733,31 +733,31 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 3, plain_global_ptr<int>, opt_striped>( sycl::sub_group, span<int, 3>, plain_global_ptr<int>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.6") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.16") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META94:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META97:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META95:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META98:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 3 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM3EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM3EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP100:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP101:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm3EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -765,31 +765,31 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 16, plain_global_ptr<int>, opt_striped>( sycl::sub_group, span<int, 16>, plain_global_ptr<int>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm16EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.15") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm16EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.29") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META101:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META104:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META102:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META105:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 16 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM16EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM16EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP107:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm16EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP108:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm16EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void @@ -797,30 +797,30 @@ template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< template SYCL_EXTERNAL void sycl::ext::oneapi::experimental::group_store< sycl::sub_group, int, 11, plain_global_ptr<int>, opt_striped>( sycl::sub_group, span<int, 11>, plain_global_ptr<int>, opt_striped); -// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm11EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESO_NS0_4spanISM_XT1_EEESN_T3_( -// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.16") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.11") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META15]] !sycl_fixed_targets [[META6]] { +// CHECK-LABEL: define weak_odr dso_local spir_func void @_ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm11EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESP_NS0_4spanISN_XT1_EEESO_T3_( +// CHECK-SAME: ptr noundef byval(%"struct.sycl::_V1::sub_group") align 1 [[G:%.*]], ptr noundef byval(%"class.sycl::_V1::span.30") align 8 [[IN:%.*]], ptr addrspace(1) noundef [[OUT_PTR:%.*]], ptr noundef byval(%"class.sycl::_V1::ext::oneapi::experimental::properties.25") align 1 [[PROPS:%.*]]) local_unnamed_addr #[[ATTR0]] comdat !srcloc [[META16]] !sycl_fixed_targets [[META7]] { // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA11]] +// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr [[IN]], align 8, !tbaa [[TBAA12]] // CHECK-NEXT: [[TMP1:%.*]] = inttoptr i64 [[TMP0]] to ptr addrspace(4) // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] -// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA7]], !noalias [[META108:![0-9]+]] -// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA7]], !noalias [[META111:![0-9]+]] +// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupLocalInvocationId, align 4, !tbaa [[TBAA8]], !noalias [[META109:![0-9]+]] +// CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr addrspace(1) @__spirv_BuiltInSubgroupSize, align 4, !tbaa [[TBAA8]], !noalias [[META112:![0-9]+]] // CHECK-NEXT: br label [[FOR_COND_I:%.*]] // CHECK: for.cond.i: // CHECK-NEXT: [[I_0_I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC_I:%.*]], [[FOR_BODY_I:%.*]] ] // CHECK-NEXT: [[CMP_I:%.*]] = icmp samesign ult i32 [[I_0_I]], 11 -// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM11EPU3AS1INS3_10PROPERTIESIST5TUPLEIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSA_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSA_INS3_14FULL_GROUP_KEYEJEEENSA_INS3_6DETAIL9NAIVE_KEYEJEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] +// CHECK-NEXT: br i1 [[CMP_I]], label [[FOR_BODY_I]], label [[_ZN4SYCL3_V13EXT6ONEAPI12EXPERIMENTAL11GROUP_STOREINS0_9SUB_GROUPEILM11EPU3AS1INS3_10PROPERTIESINS3_6DETAIL20PROPERTIES_TYPE_LISTIJNS3_14PROPERTY_VALUEINS3_18DATA_PLACEMENT_KEYEJST17INTEGRAL_CONSTANTIILI1EEEEENSB_INS3_21CONTIGUOUS_MEMORY_KEYEJEEENSB_INS3_14FULL_GROUP_KEYEJEEENSB_INS9_9NAIVE_KEYEJEEEEEEEEEENST9ENABLE_IFIXAASR6DETAILE18VERIFY_STORE_TYPESIT0_T2_ESR6DETAILE18IS_GENERIC_GROUP_VIT_EEVE4TYPEESR_NS0_4SPANISP_XT1_EEESQ_T3__EXIT:%.*]] // CHECK: for.body.i: // CHECK-NEXT: [[CONV_I:%.*]] = zext nneg i32 [[I_0_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I_I:%.*]] = getelementptr inbounds nuw i32, ptr addrspace(4) [[TMP1]], i64 [[CONV_I]] -// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspace(4) [[ARRAYIDX_I_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[MUL_I_I:%.*]] = mul i32 [[TMP3]], [[I_0_I]] // CHECK-NEXT: [[ADD_I_I:%.*]] = add i32 [[TMP2]], [[MUL_I_I]] // CHECK-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[ADD_I_I]] to i64 // CHECK-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i32, ptr addrspace(1) [[OUT_PTR]], i64 [[IDXPROM_I]] -// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA7]] +// CHECK-NEXT: store i32 [[TMP4]], ptr addrspace(1) [[ARRAYIDX_I]], align 4, !tbaa [[TBAA8]] // CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[I_0_I]], 1 -// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP114:![0-9]+]] -// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm11EPU3AS1iNS3_10propertiesISt5tupleIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSA_INS3_21contiguous_memory_keyEJEEENSA_INS3_14full_group_keyEJEEENSA_INS3_6detail9naive_keyEJEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: +// CHECK-NEXT: br label [[FOR_COND_I]], !llvm.loop [[LOOP115:![0-9]+]] +// CHECK: _ZN4sycl3_V13ext6oneapi12experimental11group_storeINS0_9sub_groupEiLm11EPU3AS1iNS3_10propertiesINS3_6detail20properties_type_listIJNS3_14property_valueINS3_18data_placement_keyEJSt17integral_constantIiLi1EEEEENSB_INS3_21contiguous_memory_keyEJEEENSB_INS3_14full_group_keyEJEEENSB_INS9_9naive_keyEJEEEEEEEEEENSt9enable_ifIXaasr6detailE18verify_store_typesIT0_T2_Esr6detailE18is_generic_group_vIT_EEvE4typeESR_NS0_4spanISP_XT1_EEESQ_T3_.exit: // CHECK-NEXT: tail call spir_func void @_Z22__spirv_ControlBarrierjjj(i32 noundef 3, i32 noundef 3, i32 noundef 912) #[[ATTR5]] // CHECK-NEXT: ret void diff --git a/sycl/test/extensions/properties/properties_ctor_negative.cpp b/sycl/test/extensions/properties/properties_ctor_negative.cpp index 5677b305acd63..e99e064944e39 100644 --- a/sycl/test/extensions/properties/properties_ctor_negative.cpp +++ b/sycl/test/extensions/properties/properties_ctor_negative.cpp @@ -5,11 +5,9 @@ #include "mock_compile_time_properties.hpp" int main() { - // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Unrecognized property in property list.}} - // expected-error@+1 {{no viable constructor or deduction guide for deduction of template arguments of 'sycl::ext::oneapi::experimental::properties'}} + // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}} Non-property argument!}} auto InvalidPropertyList1 = sycl::ext::oneapi::experimental::properties(1); - // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Unrecognized property in property list.}} - // expected-error@+1 {{no viable constructor or deduction guide for deduction of template arguments of 'sycl::ext::oneapi::experimental::properties'}} + // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}} Non-property argument!}} auto InvalidPropertyList2 = sycl::ext::oneapi::experimental::properties( sycl::ext::oneapi::experimental::foo{1}, true); // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Duplicate properties in property list.}} @@ -30,59 +28,56 @@ int main() { sycl::ext::oneapi::experimental::boo<int>, sycl::ext::oneapi::experimental::bar, sycl::ext::oneapi::experimental::fir(3.14, false)); - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: One or more property argument is not a property in the property list.}} + // expected-error@+2 {{excess elements in struct initializer}} sycl::ext::oneapi::experimental::empty_properties_t InvalidPropertyList7{ sycl::ext::oneapi::experimental::foo{0}}; - /* - // TODO: Uncomment these lines once https://github.com/intel/llvm/issues/13677 - // has been fixed. + + // expected-error-re@+2 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::foo{0}}) InvalidPropertyList8{ sycl::ext::oneapi::experimental::foo{0}, sycl::ext::oneapi::experimental::foo{1}}; + // expected-error-re@+2 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::boo<int>}) InvalidPropertyList9{ sycl::ext::oneapi::experimental::boo<int>, sycl::ext::oneapi::experimental::boo<int>}; + // expected-error-re@+3 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::boo<int>, sycl::ext::oneapi::experimental::foo{0}}) InvalidPropertyList10{ sycl::ext::oneapi::experimental::boo<int>, sycl::ext::oneapi::experimental::foo{0}, sycl::ext::oneapi::experimental::foo{1}}; + // expected-error-re@+2 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::boo<int>}) InvalidPropertyList11{ sycl::ext::oneapi::experimental::foo{0}, sycl::ext::oneapi::experimental::boo<int>, sycl::ext::oneapi::experimental::boo<int>}; - */ // TODO: For the following cases, the second error could be removed by moving // the static assert out of the Extract function. However, this // currently causes the same Clang crash as above. - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Each property in the property list must either be given an argument in the constructor or be default-constructible.}} - // expected-error@sycl/ext/oneapi/properties/properties.hpp:* {{no matching constructor for initialization of 'sycl::ext::oneapi::experimental::fir'}} + // expected-error-re@+3 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties( sycl::ext::oneapi::experimental::bar, sycl::ext::oneapi::experimental::fir(3.14, false))) InvalidPropertyList12{ sycl::ext::oneapi::experimental::bar}; - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Each property in the property list must either be given an argument in the constructor or be default-constructible.}} - // expected-error@sycl/ext/oneapi/properties/properties.hpp:* {{no matching constructor for initialization of 'sycl::ext::oneapi::experimental::fir'}} + // expected-error-re@+3 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties( sycl::ext::oneapi::experimental::fir(3.14, false))) InvalidPropertyList13{}; - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Each property in the property list must either be given an argument in the constructor or be default-constructible.}} - // expected-error@sycl/ext/oneapi/properties/properties.hpp:* {{no matching constructor for initialization of 'sycl::ext::oneapi::experimental::fir'}} - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: One or more property argument is not a property in the property list.}} + // expected-error-re@+3 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties( sycl::ext::oneapi::experimental::bar, sycl::ext::oneapi::experimental::fir(3.14, false))) InvalidPropertyList14{ sycl::ext::oneapi::experimental::boo<int>, sycl::ext::oneapi::experimental::bar}; - // expected-error@+2 {{no matching constructor for initialization of }} + // expected-error-re@+2 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties( sycl::ext::oneapi::experimental::foo{1})) InvalidPropertyList15{1}; - // expected-error@+3 {{no matching constructor for initialization of }} + // expected-error-re@+3 {{no matching constructor for initialization of {{.+}}}} decltype(sycl::ext::oneapi::experimental::properties( sycl::ext::oneapi::experimental::bar, sycl::ext::oneapi::experimental::foo{1})) InvalidPropertyList16{ diff --git a/sycl/test/extensions/properties/properties_get_property_negative.cpp b/sycl/test/extensions/properties/properties_get_property_negative.cpp index 77a8d1edb35ee..affd62b395788 100644 --- a/sycl/test/extensions/properties/properties_get_property_negative.cpp +++ b/sycl/test/extensions/properties/properties_get_property_negative.cpp @@ -6,12 +6,10 @@ int main() { auto EmptyPropertyList = sycl::ext::oneapi::experimental::properties(); - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Property list does not contain the requested property.}} - // expected-error-re@+1 {{variable has incomplete type {{.+}}}} + // expected-error@+1 {{no matching function for call to 'get_property'}} constexpr auto boo_val1 = decltype(EmptyPropertyList)::get_property< sycl::ext::oneapi::experimental::boo_key>(); - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Property list does not contain the requested property.}} - // expected-error-re@+1 {{no viable conversion from {{.+}} to 'sycl::ext::oneapi::experimental::foo'}} + // expected-error@+3 {{no matching member function for call to 'get_property'}} sycl::ext::oneapi::experimental::foo foo_val1 = EmptyPropertyList .get_property<sycl::ext::oneapi::experimental::foo_key>(); @@ -20,12 +18,10 @@ int main() { auto PopulatedPropertyList = sycl::ext::oneapi::experimental::properties( sycl::ext::oneapi::experimental::foz{.0f, true}, sycl::ext::oneapi::experimental::bar); - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Property list does not contain the requested property.}} - // expected-error-re@+1 {{variable has incomplete type {{.+}}}} + // expected-error@+1 {{no matching function for call to 'get_property'}} constexpr auto boo_val2 = decltype(PopulatedPropertyList)::get_property< sycl::ext::oneapi::experimental::boo_key>(); - // expected-error-re@sycl/ext/oneapi/properties/properties.hpp:* {{static assertion failed due to requirement {{.+}}: Property list does not contain the requested property.}} - // expected-error-re@+1 {{no viable conversion from {{.+}} to 'sycl::ext::oneapi::experimental::foo'}} + // expected-error@+3 {{no matching member function for call to 'get_property'}} sycl::ext::oneapi::experimental::foo foo_val2 = PopulatedPropertyList .get_property<sycl::ext::oneapi::experimental::foo_key>(); diff --git a/sycl/test/extensions/properties/properties_kernel_negative.cpp b/sycl/test/extensions/properties/properties_kernel_negative.cpp index 52cbba23e50eb..5b00fa954fd7e 100644 --- a/sycl/test/extensions/properties/properties_kernel_negative.cpp +++ b/sycl/test/extensions/properties/properties_kernel_negative.cpp @@ -84,77 +84,77 @@ void check_work_group_size() { // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} // expected-error-re@sycl/handler.hpp:* {{static assertion failed due to requirement {{.+}}: Template type is not a property list.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1>}, KernelFunctorWithWGSize<2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<1, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1>}, KernelFunctorWithWGSize<1, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<2, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1>}, KernelFunctorWithWGSize<2, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<2, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1>}, KernelFunctorWithWGSize<2, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<1, 1, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1, 1>}, KernelFunctorWithWGSize<1, 1, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<1, 2, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1, 1>}, KernelFunctorWithWGSize<1, 2, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<2, 1, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1, 1>}, KernelFunctorWithWGSize<2, 1, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<1, 2, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1, 1>}, KernelFunctorWithWGSize<1, 2, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<2, 2, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1, 1>}, KernelFunctorWithWGSize<2, 2, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<2, 1, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1, 1>}, KernelFunctorWithWGSize<2, 1, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSize<2, 2, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size<1, 1, 1>}, @@ -221,77 +221,77 @@ void check_work_group_size_hint() { sycl::queue Q; // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1>}, KernelFunctorWithWGSizeHint<2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<1, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1>}, KernelFunctorWithWGSizeHint<1, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<2, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1>}, KernelFunctorWithWGSizeHint<2, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<2, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1>}, KernelFunctorWithWGSizeHint<2, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<1, 1, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1, 1>}, KernelFunctorWithWGSizeHint<1, 1, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<1, 2, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1, 1>}, KernelFunctorWithWGSizeHint<1, 2, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<2, 1, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1, 1>}, KernelFunctorWithWGSizeHint<2, 1, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<1, 2, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1, 1>}, KernelFunctorWithWGSizeHint<1, 2, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<2, 2, 1>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1, 1>}, KernelFunctorWithWGSizeHint<2, 2, 1>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<2, 1, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1, 1>}, KernelFunctorWithWGSizeHint<2, 1, 2>{}); // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note-re@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithWGSizeHint<2, 2, 2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::work_group_size_hint_key, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>, std::integral_constant<unsigned long{{( long)?}}, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::work_group_size_hint<1, 1, 1>}, @@ -310,7 +310,7 @@ void check_sub_group_size() { sycl::queue Q; // expected-error-re@sycl/ext/oneapi/properties/property_utils.hpp:* {{static assertion failed due to requirement {{.+}}: Failed to merge property lists due to conflicting properties.}} - // expected-note@+1 {{in instantiation of function template specialization 'sycl::queue::single_task<sycl::detail::auto_name, KernelFunctorWithSGSize<2>, sycl::ext::oneapi::experimental::properties<std::tuple<sycl::ext::oneapi::experimental::property_value<sycl::ext::oneapi::experimental::sub_group_size_key, std::integral_constant<unsigned int, 1>>>>>' requested here}} + // expected-note-re@+1 {{in instantiation of function template specialization {{.+}}}} Q.single_task( sycl::ext::oneapi::experimental::properties{ sycl::ext::oneapi::experimental::sub_group_size<1>},