packages/glibc
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

!979 [sync] PR-978: x86: Set preferred CPU features and default NT threshold for Zhaoxin processors

Browse Source 
From: @openeuler-sync-bot 
Reviewed-by: @liqingqing_1229 
Signed-off-by: @liqingqing_1229
This commit is contained in:
openeuler-ci-bot 2025-01-20 06:17:40 +00:00 committed by Gitee
commit 8b46cd9a47
No known key found for this signature in database
GPG Key ID: 173E9B9CA92EEF8F
4 changed files with 237 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

102
0001-x86-Set-preferred-CPU-features-on-the-KH-40000-and-K.patch Normal file
View File

@ -0,0 +1,102 @@
From 5f5b877974cecf892346ae534edc4db9e8fbc75b Mon Sep 17 00:00:00 2001
From: May <mayshao-oc@zhaoxin.com>
Date: Wed, 15 Jan 2025 10:25:48 +0800
Subject: [PATCH 1/3] x86:Set preferred CPU features on the KH-40000 and
KX-7000 Zhaoxin processors
Fix code formatting under the Zhaoxin branch and add comments for
different Zhaoxin models.
Unaligned AVX load are slower on KH-40000 and KX-7000, so disable
the AVX_Fast_Unaligned_Load.
Enable Prefer_No_VZEROUPPER and Fast_Unaligned_Load features to
use sse2_unaligned version of memset,strcpy and strcat.
Signed-off-by: May <mayshao-oc@zhaoxin.com>
---
sysdeps/x86/cpu-features.c | 51 ++++++++++++++++++++++++++------------
1 file changed, 35 insertions(+), 16 deletions(-)
diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c
index badf0888..43b5f562 100644
--- a/sysdeps/x86/cpu-features.c
+++ b/sysdeps/x86/cpu-features.c
@@ -907,39 +907,58 @@ https://www.intel.com/content/www/us/en/support/articles/000059422/processors.ht
model += extended_model;
if (family == 0x6)
- {
- if (model == 0xf || model == 0x19)
- {
+ {
+ /* Tuning for older Zhaoxin processors. */
+ if (model == 0xf || model == 0x19)
+ {
CPU_FEATURE_UNSET (cpu_features, AVX);
CPU_FEATURE_UNSET (cpu_features, AVX2);
- cpu_features->preferred[index_arch_Slow_SSE4_2]
- |= bit_arch_Slow_SSE4_2;
+ cpu_features->preferred[index_arch_Slow_SSE4_2]
+ |= bit_arch_Slow_SSE4_2;
+ /* Unaligned AVX loads are slower. */
cpu_features->preferred[index_arch_AVX_Fast_Unaligned_Load]
- &= ~bit_arch_AVX_Fast_Unaligned_Load;
- }
- }
+ &= ~bit_arch_AVX_Fast_Unaligned_Load;
+ }
+ }
else if (family == 0x7)
- {
- if (model == 0x1b)
+ {
+ switch (model)
{
+ /* Wudaokou microarch tuning. */
+ case 0x1b:
CPU_FEATURE_UNSET (cpu_features, AVX);
CPU_FEATURE_UNSET (cpu_features, AVX2);
cpu_features->preferred[index_arch_Slow_SSE4_2]
- |= bit_arch_Slow_SSE4_2;
+ |= bit_arch_Slow_SSE4_2;
cpu_features->preferred[index_arch_AVX_Fast_Unaligned_Load]
- &= ~bit_arch_AVX_Fast_Unaligned_Load;
- }
- else if (model == 0x3b)
- {
+ &= ~bit_arch_AVX_Fast_Unaligned_Load;
+ break;
+
+ /* Lujiazui microarch tuning. */
+ case 0x3b:
CPU_FEATURE_UNSET (cpu_features, AVX);
CPU_FEATURE_UNSET (cpu_features, AVX2);
cpu_features->preferred[index_arch_AVX_Fast_Unaligned_Load]
- &= ~bit_arch_AVX_Fast_Unaligned_Load;
+ &= ~bit_arch_AVX_Fast_Unaligned_Load;
+ break;
+
+ /* Yongfeng and Shijidadao mircoarch tuning. */
+ case 0x5b:
+ case 0x6b:
+ cpu_features->preferred[index_arch_AVX_Fast_Unaligned_Load]
+ &= ~bit_arch_AVX_Fast_Unaligned_Load;
+
+ /* To use sse2_unaligned versions of memset, strcpy and strcat.
+ */
+ cpu_features->preferred[index_arch_Prefer_No_VZEROUPPER]
+ |= (bit_arch_Prefer_No_VZEROUPPER
+ | bit_arch_Fast_Unaligned_Load);
+ break;
}
}
}
--
2.27.0

77
0002-x86_64-Optimize-large-size-copy-in-memmove-ssse3.patch Normal file
View File

@ -0,0 +1,77 @@
From 59b19d50bd70c08e5c9f5db1742600b7b76df94a Mon Sep 17 00:00:00 2001
From: May <mayshao-oc@zhaoxin.com>
Date: Wed, 15 Jan 2025 10:32:17 +0800
Subject: [PATCH 2/3] x86_64: Optimize large size copy in memmove-ssse3
This patch optimizes large size copy using normal store when src > dst
and overlap. Make it the same as the logic in memmove-vec-unaligned-erms.S.
Current memmove-ssse3 use '__x86_shared_cache_size_half' as the non-
temporal threshold, this patch updates that value to
'__x86_shared_non_temporal_threshold'. Currently, the
__x86_shared_non_temporal_threshold is cpu-specific, and different CPUs
will have different values based on the related nt-benchmark results.
However, in memmove-ssse3, the nontemporal threshold uses
'__x86_shared_cache_size_half', which sounds unreasonable.
The performance is not changed drastically although shows overall
improvements without any major regressions or gains.
Results on Zhaoxin KX-7000:
bench-memcpy geometric_mean(N=20) New / Original: 0.999
bench-memcpy-random geometric_mean(N=20) New / Original: 0.999
bench-memcpy-large geometric_mean(N=20) New / Original: 0.978
bench-memmove geometric_mean(N=20) New / Original: 1.000
bench-memmmove-large geometric_mean(N=20) New / Original: 0.962
Results on Intel Core i5-6600K:
bench-memcpy geometric_mean(N=20) New / Original: 1.001
bench-memcpy-random geometric_mean(N=20) New / Original: 0.999
bench-memcpy-large geometric_mean(N=20) New / Original: 1.001
bench-memmove geometric_mean(N=20) New / Original: 0.995
bench-memmmove-large geometric_mean(N=20) New / Original: 0.936
Signed-off-by: May <mayshao-oc@zhaoxin.com>
---
sysdeps/x86_64/multiarch/memmove-ssse3.S | 14 +++++++++-----
1 file changed, 9 insertions(+), 5 deletions(-)
diff --git a/sysdeps/x86_64/multiarch/memmove-ssse3.S b/sysdeps/x86_64/multiarch/memmove-ssse3.S
index 460b0ec0..69561628 100644
--- a/sysdeps/x86_64/multiarch/memmove-ssse3.S
+++ b/sysdeps/x86_64/multiarch/memmove-ssse3.S
@@ -151,13 +151,10 @@ L(more_2x_vec):
loop. */
movups %xmm0, (%rdi)
-# ifdef SHARED_CACHE_SIZE_HALF
- cmp $SHARED_CACHE_SIZE_HALF, %RDX_LP
-# else
- cmp __x86_shared_cache_size_half(%rip), %rdx
-# endif
+ cmp __x86_shared_non_temporal_threshold(%rip), %rdx
ja L(large_memcpy)
+L(loop_fwd):
leaq -64(%rdi, %rdx), %r8
andq $-16, %rdi
movl $48, %edx
@@ -199,6 +196,13 @@ L(large_memcpy):
movups -64(%r9, %rdx), %xmm10
movups -80(%r9, %rdx), %xmm11
+ /* Check if src and dst overlap. If they do use cacheable
+ writes to potentially gain positive interference between
+ the loads during the memmove. */
+ subq %rdi, %r9
+ cmpq %rdx, %r9
+ jb L(loop_fwd)
+
sall $5, %ecx
leal (%rcx, %rcx, 2), %r8d
leaq -96(%rdi, %rdx), %rcx
--
2.27.0

50
0003-x86-Set-default-non_temporal_threshold-for-Zhaoxin-p.patch Normal file
View File

@ -0,0 +1,50 @@
From 6374dceef2bf72456144184cb70bdd216a744d61 Mon Sep 17 00:00:00 2001
From: May <mayshao-oc@zhaoxin.com>
Date: Wed, 15 Jan 2025 10:42:33 +0800
Subject: [PATCH 3/3] x86: Set default non_temporal_threshold for Zhaoxin
processors
Current 'non_temporal_threshold' set to 'non_temporal_threshold_lowbound'
on Zhaoxin processors without ERMS. The default
'non_temporal_threshold_lowbound' is too small for the KH-40000 and KX-7000
Zhaoxin processors, this patch updates the value to
'shared / cachesize_non_temporal_divisor'.
Signed-off-by: May <mayshao-oc@zhaoxin.com>
---
sysdeps/x86/cpu-features.c | 1 +
sysdeps/x86/dl-cacheinfo.h | 6 ++++--
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/sysdeps/x86/cpu-features.c b/sysdeps/x86/cpu-features.c
index 43b5f562..f752ebd2 100644
--- a/sysdeps/x86/cpu-features.c
+++ b/sysdeps/x86/cpu-features.c
@@ -949,6 +949,7 @@ https://www.intel.com/content/www/us/en/support/articles/000059422/processors.ht
/* Yongfeng and Shijidadao mircoarch tuning. */
case 0x5b:
+ cpu_features->cachesize_non_temporal_divisor = 2;
case 0x6b:
cpu_features->preferred[index_arch_AVX_Fast_Unaligned_Load]
&= ~bit_arch_AVX_Fast_Unaligned_Load;
diff --git a/sysdeps/x86/dl-cacheinfo.h b/sysdeps/x86/dl-cacheinfo.h
index 6c774042..bd2a9122 100644
--- a/sysdeps/x86/dl-cacheinfo.h
+++ b/sysdeps/x86/dl-cacheinfo.h
@@ -940,8 +940,10 @@ dl_init_cacheinfo (struct cpu_features *cpu_features)
/* If no ERMS, we use the per-thread L3 chunking. Normal cacheable stores run
a higher risk of actually thrashing the cache as they don't have a HW LRU
hint. As well, their performance in highly parallel situations is
- noticeably worse. */
- if (!CPU_FEATURE_USABLE_P (cpu_features, ERMS))
+ noticeably worse. Zhaoxin processors are an exception, the lowbound is not
+ suitable for them based on actual test data. */
+ if (!CPU_FEATURE_USABLE_P (cpu_features, ERMS)
+ && cpu_features->basic.kind != arch_kind_zhaoxin)
non_temporal_threshold = non_temporal_threshold_lowbound;
/* SIZE_MAX >> 4 because memmove-vec-unaligned-erms right-shifts the value of
'x86_non_temporal_threshold' by `LOG_4X_MEMCPY_THRESH` (4) and it is best
--
2.27.0

9
glibc.spec
View File

@ -67,7 +67,7 @@
############################################################################## ##############################################################################
Name: glibc Name: glibc
Version: 2.38 Version: 2.38
Release: 50 Release: 51
Summary: The GNU libc libraries Summary: The GNU libc libraries
License: %{all_license} License: %{all_license}
URL: http://www.gnu.org/software/glibc/ URL: http://www.gnu.org/software/glibc/
@ -291,6 +291,10 @@ Patch9031: 0021-Sw64-Add-test_numdouble.h-and-test_numfloat.h.patch
Patch9032: 0022-Sw64-Fix-posix-tst-glob_lstat_compat-on-sw64.patch Patch9032: 0022-Sw64-Fix-posix-tst-glob_lstat_compat-on-sw64.patch
Patch9033: 0023-Sw64-add-getopt-weak-alias.patch Patch9033: 0023-Sw64-add-getopt-weak-alias.patch
Patch9034: 0001-x86-Set-preferred-CPU-features-on-the-KH-40000-and-K.patch
Patch9035: 0002-x86_64-Optimize-large-size-copy-in-memmove-ssse3.patch
Patch9036: 0003-x86-Set-default-non_temporal_threshold-for-Zhaoxin-p.patch
Provides: ldconfig rtld(GNU_HASH) bundled(gnulib) Provides: ldconfig rtld(GNU_HASH) bundled(gnulib)
BuildRequires: audit-libs-devel >= 1.1.3, sed >= 3.95, libcap-devel, gettext BuildRequires: audit-libs-devel >= 1.1.3, sed >= 3.95, libcap-devel, gettext
@ -1467,6 +1471,9 @@ fi
%endif %endif
%changelog %changelog
* Wed Jan 15 2025 MayShao <mayshao-oc@zhaoxin.com> - 2.38-51
- x86: Set preferred CPU features and default NT threshold for Zhaoxin processors
* Tue Jan 07 2025 Peng Fan <fanpeng@loongson.cn> - 2.38-50 * Tue Jan 07 2025 Peng Fan <fanpeng@loongson.cn> - 2.38-50
- LoongArch: Force SHMLBA the same as kernel - LoongArch: Force SHMLBA the same as kernel