Merge branch 'main' into Getting_started_with_Trixi

docs/literate/src/files/index.jl

@@ -115,20 +115,26 @@
 # software in the Trixi.jl ecosystem, and then run a simulation using Trixi.jl on said mesh.
 # In the end, the tutorial briefly explains how to simulate an example using AMR via `P4estMesh`.
 
-# ### [16 Explicit time stepping](@ref time_stepping)
+# ### [16 P4est mesh from gmsh](@ref p4est_from_gmsh)
+#-
+# This tutorial describes how to obtain a [`P4estMesh`](@ref) from an existing mesh generated
+# by [`gmsh`](https://gmsh.info/) or any other meshing software that can export to the Abaqus
+# input `.inp` format. The tutorial demonstrates how edges/faces can be associated with boundary conditions based on the physical nodesets.
+
+# ### [17 Explicit time stepping](@ref time_stepping)
 #-
 # This tutorial is about time integration using [OrdinaryDiffEq.jl](https://github.com/SciML/OrdinaryDiffEq.jl).
 # It explains how to use their algorithms and presents two types of time step choices - with error-based
 # and CFL-based adaptive step size control.
 
-# ### [17 Differentiable programming](@ref differentiable_programming)
+# ### [18 Differentiable programming](@ref differentiable_programming)
 #-
 # This part deals with some basic differentiable programming topics. For example, a Jacobian, its
 # eigenvalues and a curve of total energy (through the simulation) are calculated and plotted for
 # a few semidiscretizations. Moreover, we calculate an example for propagating errors with Measurement.jl
 # at the end.
 
-# ### [18 Custom semidiscretization](@ref custom_semidiscretization)
+# ### [19 Custom semidiscretization](@ref custom_semidiscretization)
 #-
 # This tutorial describes the [semidiscretiations](@ref overview-semidiscretizations) of Trixi.jl
 # and explains how to extend them for custom tasks.

docs/src/performance.md

@@ -267,3 +267,14 @@ requires. It can thus be seen as a proxy for "energy used" and, as an extension,
     timing result, you need to set the analysis interval such that the
     `AnalysisCallback` is invoked at least once during the course of the simulation and
     discard the first PID value.
+
+## Performance issues with multi-threaded reductions
+[False sharing](https://en.wikipedia.org/wiki/False_sharing) is a known performance issue
+for systems with distributed caches. It also occurred for the implementation of a thread
+parallel bounds checking routine for the subcell IDP limiting
+in [PR #1736](https://github.com/trixi-framework/Trixi.jl/pull/1736).
+After some [testing and discussion](https://github.com/trixi-framework/Trixi.jl/pull/1736#discussion_r1423881895),
+it turned out that initializing a vector of length `n * Threads.nthreads()` and only using every
+n-th entry instead of a vector of length `Threads.nthreads()` fixes the problem.
+Since there are no processors with caches over 128B, we use `n = 128B / size(uEltype)`.
+Now, the bounds checking routine of the IDP limiting scales as hoped.