public inbox for glibc-cvs@sourceware.org
help / color / mirror / Atom feed
* [glibc/release/2.33/master] x86: Set Prefer_No_VZEROUPPER and add Prefer_AVX2_STRCMP
@ 2022-01-27 19:32 H.J. Lu
0 siblings, 0 replies; only message in thread
From: H.J. Lu @ 2022-01-27 19:32 UTC (permalink / raw)
To: glibc-cvs
https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=f0a2b671471bf1fe943feb4a2d00409625e05ba4
commit f0a2b671471bf1fe943feb4a2d00409625e05ba4
Author: H.J. Lu <hjl.tools@gmail.com>
Date: Fri Feb 26 05:36:59 2021 -0800
x86: Set Prefer_No_VZEROUPPER and add Prefer_AVX2_STRCMP
1. Set Prefer_No_VZEROUPPER if RTM is usable to avoid RTM abort triggered
by VZEROUPPER inside a transactionally executing RTM region.
2. Since to compare 2 32-byte strings, 256-bit EVEX strcmp requires 2
loads, 3 VPCMPs and 2 KORDs while AVX2 strcmp requires 1 load, 2 VPCMPEQs,
1 VPMINU and 1 VPMOVMSKB, AVX2 strcmp is faster than EVEX strcmp. Add
Prefer_AVX2_STRCMP to prefer AVX2 strcmp family functions.
(cherry picked from commit 1da50d4bda07f04135dca39f40e79fc9eabed1f8)
Diff:
---
sysdeps/x86/cpu-features.c | 20 ++++++++++++++++++--
sysdeps/x86/cpu-tunables.c | 2 ++
.../cpu-features-preferred_feature_index_1.def | 1 +
3 files changed, 21 insertions(+), 2 deletions(-)
diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c
index 73b0a4dc9a..dfad82f14c 100644
--- a/sysdeps/x86/cpu-features.c
+++ b/sysdeps/x86/cpu-features.c
@@ -520,8 +520,24 @@ init_cpu_features (struct cpu_features *cpu_features)
cpu_features->preferred[index_arch_Prefer_No_VZEROUPPER]
|= bit_arch_Prefer_No_VZEROUPPER;
else
- cpu_features->preferred[index_arch_Prefer_No_AVX512]
- |= bit_arch_Prefer_No_AVX512;
+ {
+ cpu_features->preferred[index_arch_Prefer_No_AVX512]
+ |= bit_arch_Prefer_No_AVX512;
+
+ /* Avoid RTM abort triggered by VZEROUPPER inside a
+ transactionally executing RTM region. */
+ if (CPU_FEATURE_USABLE_P (cpu_features, RTM))
+ cpu_features->preferred[index_arch_Prefer_No_VZEROUPPER]
+ |= bit_arch_Prefer_No_VZEROUPPER;
+
+ /* Since to compare 2 32-byte strings, 256-bit EVEX strcmp
+ requires 2 loads, 3 VPCMPs and 2 KORDs while AVX2 strcmp
+ requires 1 load, 2 VPCMPEQs, 1 VPMINU and 1 VPMOVMSKB,
+ AVX2 strcmp is faster than EVEX strcmp. */
+ if (CPU_FEATURE_USABLE_P (cpu_features, AVX2))
+ cpu_features->preferred[index_arch_Prefer_AVX2_STRCMP]
+ |= bit_arch_Prefer_AVX2_STRCMP;
+ }
}
/* This spells out "AuthenticAMD" or "HygonGenuine". */
else if ((ebx == 0x68747541 && ecx == 0x444d4163 && edx == 0x69746e65)
diff --git a/sysdeps/x86/cpu-tunables.c b/sysdeps/x86/cpu-tunables.c
index 126896f41b..a90df39b78 100644
--- a/sysdeps/x86/cpu-tunables.c
+++ b/sysdeps/x86/cpu-tunables.c
@@ -238,6 +238,8 @@ TUNABLE_CALLBACK (set_hwcaps) (tunable_val_t *valp)
CHECK_GLIBC_IFUNC_PREFERRED_BOTH (n, cpu_features,
Fast_Copy_Backward,
disable, 18);
+ CHECK_GLIBC_IFUNC_PREFERRED_NEED_BOTH
+ (n, cpu_features, Prefer_AVX2_STRCMP, AVX2, disable, 18);
}
break;
case 19:
diff --git a/sysdeps/x86/include/cpu-features-preferred_feature_index_1.def b/sysdeps/x86/include/cpu-features-preferred_feature_index_1.def
index 06af1a8dd5..133aab19f1 100644
--- a/sysdeps/x86/include/cpu-features-preferred_feature_index_1.def
+++ b/sysdeps/x86/include/cpu-features-preferred_feature_index_1.def
@@ -32,3 +32,4 @@ BIT (Prefer_ERMS)
BIT (Prefer_No_AVX512)
BIT (MathVec_Prefer_No_AVX512)
BIT (Prefer_FSRM)
+BIT (Prefer_AVX2_STRCMP)
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2022-01-27 19:32 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-01-27 19:32 [glibc/release/2.33/master] x86: Set Prefer_No_VZEROUPPER and add Prefer_AVX2_STRCMP H.J. Lu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).