This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
[PATCH] x86: Add MathVec_Prefer_No_AVX512 to cpu-features [BZ #21967]
- From: "H.J. Lu" <hongjiu dot lu at intel dot com>
- To: GNU C Library <libc-alpha at sourceware dot org>
- Cc: Andrew Senkevich <andrew dot senkevich at intel dot com>
- Date: Fri, 1 Sep 2017 10:40:34 -0700
- Subject: [PATCH] x86: Add MathVec_Prefer_No_AVX512 to cpu-features [BZ #21967]
- Authentication-results: sourceware.org; auth=none
- Reply-to: "H.J. Lu" <hjl dot tools at gmail dot com>
AVX512 functions in mathvec are used on machines with AVX512. An AVX2
wrapper is also provided and it can be used when the AVX512 version
isn't profitable. MathVec_Prefer_No_AVX512 is addded to cpu-features.
If glibc.tune.hwcaps=MathVec_Prefer_No_AVX512 is set in GLIBC_TUNABLES
environment variable, the AVX2 wrapper will be used.
Tested on x86-64 machines with and without AVX512. Also verified
glibc.tune.hwcaps=MathVec_Prefer_No_AVX512 on AVX512 machine.
Any comments?
H.J.
---
[BZ #21967]
* sysdeps/x86/cpu-features.h (bit_arch_MathVec_Prefer_No_AVX512):
New.
(index_arch_MathVec_Prefer_No_AVX512): Likewise.
* sysdeps/x86/cpu-tunables.c (TUNABLE_CALLBACK (set_hwcaps)):
Handle MathVec_Prefer_No_AVX512.
---
sysdeps/x86/cpu-features.h | 2 ++
sysdeps/x86/cpu-tunables.c | 7 +++++++
sysdeps/x86_64/fpu/multiarch/ifunc-mathvec-avx512.h | 13 ++++++++-----
3 files changed, 17 insertions(+), 5 deletions(-)
diff --git a/sysdeps/x86/cpu-features.h b/sysdeps/x86/cpu-features.h
index 9e01781424..a032a2e168 100644
--- a/sysdeps/x86/cpu-features.h
+++ b/sysdeps/x86/cpu-features.h
@@ -40,6 +40,7 @@
#define bit_arch_Use_dl_runtime_resolve_opt (1 << 20)
#define bit_arch_Use_dl_runtime_resolve_slow (1 << 21)
#define bit_arch_Prefer_No_AVX512 (1 << 22)
+#define bit_arch_MathVec_Prefer_No_AVX512 (1 << 23)
/* CPUID Feature flags. */
@@ -239,6 +240,7 @@ extern const struct cpu_features *__get_cpu_features (void)
# define index_arch_Use_dl_runtime_resolve_opt FEATURE_INDEX_1
# define index_arch_Use_dl_runtime_resolve_slow FEATURE_INDEX_1
# define index_arch_Prefer_No_AVX512 FEATURE_INDEX_1
+# define index_arch_MathVec_Prefer_No_AVX512 FEATURE_INDEX_1
#endif /* !__ASSEMBLER__ */
diff --git a/sysdeps/x86/cpu-tunables.c b/sysdeps/x86/cpu-tunables.c
index 0ab708cca8..ec72d86f08 100644
--- a/sysdeps/x86/cpu-tunables.c
+++ b/sysdeps/x86/cpu-tunables.c
@@ -303,6 +303,13 @@ TUNABLE_CALLBACK (set_hwcaps) (tunable_val_t *valp)
disable, 23);
}
break;
+ case 24:
+ {
+ CHECK_GLIBC_IFUNC_ARCH_NEED_ARCH_BOTH
+ (n, cpu_features, MathVec_Prefer_No_AVX512,
+ AVX512F_Usable, disable, 24);
+ }
+ break;
case 26:
{
CHECK_GLIBC_IFUNC_ARCH_NEED_CPU_BOTH
diff --git a/sysdeps/x86_64/fpu/multiarch/ifunc-mathvec-avx512.h b/sysdeps/x86_64/fpu/multiarch/ifunc-mathvec-avx512.h
index 1857e1f760..fffc9da114 100644
--- a/sysdeps/x86_64/fpu/multiarch/ifunc-mathvec-avx512.h
+++ b/sysdeps/x86_64/fpu/multiarch/ifunc-mathvec-avx512.h
@@ -32,11 +32,14 @@ IFUNC_SELECTOR (void)
{
const struct cpu_features* cpu_features = __get_cpu_features ();
- if (CPU_FEATURES_ARCH_P (cpu_features, AVX512DQ_Usable))
- return OPTIMIZE (skx);
-
- if (CPU_FEATURES_ARCH_P (cpu_features, AVX512F_Usable))
- return OPTIMIZE (knl);
+ if (!CPU_FEATURES_ARCH_P (cpu_features, MathVec_Prefer_No_AVX512))
+ {
+ if (CPU_FEATURES_ARCH_P (cpu_features, AVX512DQ_Usable))
+ return OPTIMIZE (skx);
+
+ if (CPU_FEATURES_ARCH_P (cpu_features, AVX512F_Usable))
+ return OPTIMIZE (knl);
+ }
return OPTIMIZE (avx2_wrapper);
}
--
2.13.5