Force-disable BMI2 instruction set when building snappy.lib #47
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mangini
changed the title
bcny-fork-syncer
bot
force-pushed
the
compnerd/swift
branch
from
6171d35
to
1b37356
Compare
mangini
force-pushed
the
mangini/leveldb-without-bmi2
branch
2 times, most recently
from
7df18e6
to
e567284
Compare
mangini
changed the title
bcny-fork-syncer
bot
force-pushed
the
compnerd/swift
branch
from
1b37356
to
43e9ddc
Compare
mangini
force-pushed
the
mangini/leveldb-without-bmi2
branch
from
a46c953
to
ec73e88
Compare
compnerd
reviewed
Aug 6, 2024
There was a problem hiding this comment.
I'm fine with this approach, but this will need to be fast deprecated. BMI2 is not really "optional" at this point. This extension has been around for 11 years and is part of Haswell family. We are already dependent on the previous generation (Nehalem) for cx16. Given that there has been discussions for upgrading to Nehalem and Swift is going to drop support for Windows 10 with it being EOL'ed 10/14/24, this might be a tease.
…ix check for amd64
compnerd
approved these changes
Aug 6, 2024
mangini
added a commit
that referenced
this pull request
Aug 6, 2024
Firebase-cpp-sdk on Windows pulls in firebase-ios-sdk (for firestore code), which pulls in leveldb (for key/value persistence) and Snappy (for efficient data compression). For faster compression, Snappy uses one specific CPU instruction for bit manipulation, bzhi, from the BMI2 instruction set. However, this instruction crashes firestore on a CPU that does not support the BMI2 instructions, which is the case of older CPUs, for example Intel pre-Haswell and AMD pre-Excavator CPUs. By default, Snappy detects if it should use the BMI2 optimized instruction by performing a compilation test on the cmake configure phase , which obviously returns if the build machine supports BMI2, not the target machines. Snappy build setup does not support a compiler directive that overrides that specific compilation test, so it can only build for the host machine BMI2 setup. Unfortunately, due to the very involved dependency chain, there is no good way to patch Snappy's code and the easiest solution in terms of maintenance is to patch the config.h created during the configure step, which is what this PR does. It only changes the automated build process (Github Actions workflow).
bcny-fork-syncer bot
pushed a commit
that referenced
this pull request
Aug 21, 2024
Firebase-cpp-sdk on Windows pulls in firebase-ios-sdk (for firestore code), which pulls in leveldb (for key/value persistence) and Snappy (for efficient data compression). For faster compression, Snappy uses one specific CPU instruction for bit manipulation, bzhi, from the BMI2 instruction set. However, this instruction crashes firestore on a CPU that does not support the BMI2 instructions, which is the case of older CPUs, for example Intel pre-Haswell and AMD pre-Excavator CPUs. By default, Snappy detects if it should use the BMI2 optimized instruction by performing a compilation test on the cmake configure phase , which obviously returns if the build machine supports BMI2, not the target machines. Snappy build setup does not support a compiler directive that overrides that specific compilation test, so it can only build for the host machine BMI2 setup. Unfortunately, due to the very involved dependency chain, there is no good way to patch Snappy's code and the easiest solution in terms of maintenance is to patch the config.h created during the configure step, which is what this PR does. It only changes the automated build process (Github Actions workflow).
bcny-fork-syncer bot
pushed a commit
that referenced
this pull request
Sep 7, 2024
Firebase-cpp-sdk on Windows pulls in firebase-ios-sdk (for firestore code), which pulls in leveldb (for key/value persistence) and Snappy (for efficient data compression). For faster compression, Snappy uses one specific CPU instruction for bit manipulation, bzhi, from the BMI2 instruction set. However, this instruction crashes firestore on a CPU that does not support the BMI2 instructions, which is the case of older CPUs, for example Intel pre-Haswell and AMD pre-Excavator CPUs. By default, Snappy detects if it should use the BMI2 optimized instruction by performing a compilation test on the cmake configure phase , which obviously returns if the build machine supports BMI2, not the target machines. Snappy build setup does not support a compiler directive that overrides that specific compilation test, so it can only build for the host machine BMI2 setup. Unfortunately, due to the very involved dependency chain, there is no good way to patch Snappy's code and the easiest solution in terms of maintenance is to patch the config.h created during the configure step, which is what this PR does. It only changes the automated build process (Github Actions workflow).
bcny-fork-syncer bot
pushed a commit
that referenced
this pull request
Sep 11, 2024
Firebase-cpp-sdk on Windows pulls in firebase-ios-sdk (for firestore code), which pulls in leveldb (for key/value persistence) and Snappy (for efficient data compression). For faster compression, Snappy uses one specific CPU instruction for bit manipulation, bzhi, from the BMI2 instruction set. However, this instruction crashes firestore on a CPU that does not support the BMI2 instructions, which is the case of older CPUs, for example Intel pre-Haswell and AMD pre-Excavator CPUs. By default, Snappy detects if it should use the BMI2 optimized instruction by performing a compilation test on the cmake configure phase , which obviously returns if the build machine supports BMI2, not the target machines. Snappy build setup does not support a compiler directive that overrides that specific compilation test, so it can only build for the host machine BMI2 setup. Unfortunately, due to the very involved dependency chain, there is no good way to patch Snappy's code and the easiest solution in terms of maintenance is to patch the config.h created during the configure step, which is what this PR does. It only changes the automated build process (Github Actions workflow).
bcny-fork-syncer bot
pushed a commit
that referenced
this pull request
Sep 12, 2024
Firebase-cpp-sdk on Windows pulls in firebase-ios-sdk (for firestore code), which pulls in leveldb (for key/value persistence) and Snappy (for efficient data compression). For faster compression, Snappy uses one specific CPU instruction for bit manipulation, bzhi, from the BMI2 instruction set. However, this instruction crashes firestore on a CPU that does not support the BMI2 instructions, which is the case of older CPUs, for example Intel pre-Haswell and AMD pre-Excavator CPUs. By default, Snappy detects if it should use the BMI2 optimized instruction by performing a compilation test on the cmake configure phase , which obviously returns if the build machine supports BMI2, not the target machines. Snappy build setup does not support a compiler directive that overrides that specific compilation test, so it can only build for the host machine BMI2 setup. Unfortunately, due to the very involved dependency chain, there is no good way to patch Snappy's code and the easiest solution in terms of maintenance is to patch the config.h created during the configure step, which is what this PR does. It only changes the automated build process (Github Actions workflow).
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Context
Firebase-cpp-sdk on Windows pulls in firebase-ios-sdk (for firestore code), which pulls in leveldb (for key/value persistence) and Snappy (for efficient data compression). For faster compression, Snappy uses one specific CPU instruction for bit manipulation,
bzhi, from the BMI2 instruction set. However, this crashes firestore in a CPU that does not support the BMI2 instructions, which is the case of older CPUs, for example Intel pre-Haswell and AMD pre-Excavator CPUs.How this PR works
By default, Snappy detects if it should use the BMI2 optimized instruction by performing a compilation test on the cmake configure phase , which obviously returns if the build machine supports BMI2, not the target machines.
Snappy build setup does not support a compiler directive that overrides that specific compilation test, so it can only build for the host machine BMI2 setup.
Unfortunately, due to the very involved dependency chain, there is no good way to patch Snappy's code and the easiest solution in terms of maintenance is to patch the config.h created during the configure step, which is what this PR does. It only changes the automated build process (Github Actions workflow).
Type of Change
Place an
xthe applicable box:Notes
Release Notessection ofrelease_build_files/readme.md.