[Openmp-commits] [openmp] d4a7b8d - [OpenMP][libomp] avoid spin wait and yield on arm64 macOS

[Jonathan Peyton via Openmp-commits]

Author: Daniel Douglas
Date: 2022-06-24T12:02:16-05:00
New Revision: d4a7b8de527da9f3fa02a86c131b41ca52bb878c

URL: https://github.com/llvm/llvm-project/commit/d4a7b8de527da9f3fa02a86c131b41ca52bb878c
DIFF: https://github.com/llvm/llvm-project/commit/d4a7b8de527da9f3fa02a86c131b41ca52bb878c.diff

LOG: [OpenMP][libomp] avoid spin wait and yield on arm64 macOS

This patch changes the default behavior to avoid spin waiting and
yielding. (See “Don’t Keep Threads Active And Idle” section here:
https://developer.apple.com/documentation/apple-silicon/tuning-your-code-s-performance-for-apple-silicon)

We verified using instruments traces that the changes improve scheduling
behavior on macOS.

We also collected results using EPCC schedbench
(https://github.com/LangdalP/EPCC-OpenMP-micro-benchmarks) that are
attached here that show a reduction in standard deviation and max test
run time across all scheduling types. Static scheduling sees dramatic
improvements with these changes, we see a 2-4x average runtime
improvement in the benchmark.

Differential Revision: https://reviews.llvm.org/D126510

Added: 
    

Modified: 
    openmp/runtime/src/kmp.h
    openmp/runtime/src/kmp_global.cpp
    openmp/runtime/src/kmp_runtime.cpp

Removed: 
    


################################################################################
diff  --git a/openmp/runtime/src/kmp.h b/openmp/runtime/src/kmp.h
index c90931d8a9966..ac680e32d390e 100644
--- a/openmp/runtime/src/kmp.h
+++ b/openmp/runtime/src/kmp.h
@@ -3061,6 +3061,8 @@ extern int __kmp_storage_map_verbose_specified;
 #if KMP_ARCH_X86 || KMP_ARCH_X86_64
 extern kmp_cpuinfo_t __kmp_cpuinfo;
 static inline bool __kmp_is_hybrid_cpu() { return __kmp_cpuinfo.flags.hybrid; }
+#elif KMP_OS_DARWIN && KMP_ARCH_AARCH64
+static inline bool __kmp_is_hybrid_cpu() { return true; }
 #else
 static inline bool __kmp_is_hybrid_cpu() { return false; }
 #endif

diff  --git a/openmp/runtime/src/kmp_global.cpp b/openmp/runtime/src/kmp_global.cpp
index 525873646125b..99a6e88e25f8d 100644
--- a/openmp/runtime/src/kmp_global.cpp
+++ b/openmp/runtime/src/kmp_global.cpp
@@ -425,7 +425,13 @@ int __kmp_env_consistency_check = FALSE; /* KMP_CONSISTENCY_CHECK specified? */
 // 0 = never yield;
 // 1 = always yield (default);
 // 2 = yield only if oversubscribed
+#if KMP_OS_DARWIN && KMP_ARCH_AARCH64
+// Set to 0 for environments where yield is slower
+kmp_int32 __kmp_use_yield = 0;
+#else
 kmp_int32 __kmp_use_yield = 1;
+#endif
+
 // This will be 1 if KMP_USE_YIELD environment variable was set explicitly
 kmp_int32 __kmp_use_yield_exp_set = 0;
 

diff  --git a/openmp/runtime/src/kmp_runtime.cpp b/openmp/runtime/src/kmp_runtime.cpp
index a44b4d88bc0b0..2b339fd38d63c 100644
--- a/openmp/runtime/src/kmp_runtime.cpp
+++ b/openmp/runtime/src/kmp_runtime.cpp
@@ -8300,7 +8300,7 @@ void __kmp_aux_set_library(enum library_type arg) {
     break;
   case library_throughput:
     if (__kmp_dflt_blocktime == KMP_MAX_BLOCKTIME)
-      __kmp_dflt_blocktime = 200;
+      __kmp_dflt_blocktime = KMP_DEFAULT_BLOCKTIME;
     break;
   default:
     KMP_FATAL(UnknownLibraryType, arg);