system/glibc
3
0
Fork 0
Code Issues Projects Releases Wiki Activity

x86: Use AVX2 memcpy/memset on Skylake server [BZ #21396]

Browse Source 
On Skylake server, AVX512 load/store instructions in memcpy/memset may
lead to lower CPU turbo frequency in certain situations.  Use of AVX2
in memcpy/memset has been observed to have improved overall performance
in many workloads due to the higher frequency.

Since AVX512ER is unique to Xeon Phi, this patch sets Prefer_No_AVX512
if AVX512ER isn't available so that AVX2 versions of memcpy/memset are
used on Skylake server.

	[BZ #21396]
	* sysdeps/x86/cpu-features.c (init_cpu_features): Set
	Prefer_No_AVX512 if AVX512ER isn't available.
	* sysdeps/x86/cpu-features.h (bit_arch_Prefer_No_AVX512): New.
	(index_arch_Prefer_No_AVX512): Likewise.
	* sysdeps/x86_64/multiarch/memcpy.S (__new_memcpy): Don't use
	AVX512 version if Prefer_No_AVX512 is set.
	* sysdeps/x86_64/multiarch/memcpy_chk.S (__memcpy_chk):
	Likewise.
	* sysdeps/x86_64/multiarch/memmove.S (__libc_memmove): Likewise.
	* sysdeps/x86_64/multiarch/memmove_chk.S (__memmove_chk):
	Likewise.
	* sysdeps/x86_64/multiarch/mempcpy.S (__mempcpy): Likewise.
	* sysdeps/x86_64/multiarch/mempcpy_chk.S (__mempcpy_chk):
	Likewise.
	* sysdeps/x86_64/multiarch/memset.S (memset): Likewise.
	* sysdeps/x86_64/multiarch/memset_chk.S (__memset_chk):
	Likewise.
This commit is contained in:
H.J. Lu 2017-04-18 14:01:45 -07:00
parent 1c53cb49de
commit 4cb334c4d6
11 changed files with 45 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
ChangeLog
View File

@ -1,3 +1,24 @@
2017-04-18 H.J. Lu <hongjiu.lu@intel.com>
[BZ #21396]
* sysdeps/x86/cpu-features.c (init_cpu_features): Set
Prefer_No_AVX512 if AVX512ER isn't available.
* sysdeps/x86/cpu-features.h (bit_arch_Prefer_No_AVX512): New.
(index_arch_Prefer_No_AVX512): Likewise.
* sysdeps/x86_64/multiarch/memcpy.S (__new_memcpy): Don't use
AVX512 version if Prefer_No_AVX512 is set.
* sysdeps/x86_64/multiarch/memcpy_chk.S (__memcpy_chk):
Likewise.
* sysdeps/x86_64/multiarch/memmove.S (__libc_memmove): Likewise.
* sysdeps/x86_64/multiarch/memmove_chk.S (__memmove_chk):
Likewise.
* sysdeps/x86_64/multiarch/mempcpy.S (__mempcpy): Likewise.
* sysdeps/x86_64/multiarch/mempcpy_chk.S (__mempcpy_chk):
Likewise.
* sysdeps/x86_64/multiarch/memset.S (memset): Likewise.
* sysdeps/x86_64/multiarch/memset_chk.S (__memset_chk):
Likewise.
2017-04-18 H.J. Lu <hongjiu.lu@intel.com>
* sysdeps/x86/cpu-features.c (init_cpu_features): Set

6
sysdeps/x86/cpu-features.c
View File

@ -224,10 +224,14 @@ init_cpu_features (struct cpu_features *cpu_features)
|= bit_arch_AVX_Fast_Unaligned_Load;
/* Since AVX512ER is unique to Xeon Phi, set Prefer_No_VZEROUPPER
if AVX512ER is available. */
if AVX512ER is available. Don't use AVX512 to avoid lower CPU
frequency if AVX512ER isn't available. */
if (CPU_FEATURES_CPU_P (cpu_features, AVX512ER))
cpu_features->feature[index_arch_Prefer_No_VZEROUPPER]
|= bit_arch_Prefer_No_VZEROUPPER;
else
cpu_features->feature[index_arch_Prefer_No_AVX512]
|= bit_arch_Prefer_No_AVX512;
/* To avoid SSE transition penalty, use _dl_runtime_resolve_slow.
If XGETBV suports ECX == 1, use _dl_runtime_resolve_opt. */

3
sysdeps/x86/cpu-features.h
View File

@ -39,6 +39,7 @@
#define bit_arch_Prefer_ERMS (1 << 19)
#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)
/* CPUID Feature flags. */
@ -118,6 +119,7 @@
# define index_arch_Prefer_ERMS FEATURE_INDEX_1*FEATURE_SIZE
# define index_arch_Use_dl_runtime_resolve_opt FEATURE_INDEX_1*FEATURE_SIZE
# define index_arch_Use_dl_runtime_resolve_slow FEATURE_INDEX_1*FEATURE_SIZE
# define index_arch_Prefer_No_AVX512 FEATURE_INDEX_1*FEATURE_SIZE
# if defined (_LIBC) && !IS_IN (nonlib)
@ -302,6 +304,7 @@ extern const struct cpu_features *__get_cpu_features (void)
# define index_arch_Prefer_ERMS FEATURE_INDEX_1
# 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
#endif /* !__ASSEMBLER__ */

2
sysdeps/x86_64/multiarch/memcpy.S
View File

@ -32,6 +32,8 @@ ENTRY(__new_memcpy)
lea __memcpy_erms(%rip), %RAX_LP
HAS_ARCH_FEATURE (Prefer_ERMS)
jnz 2f
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 1f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 1f
lea __memcpy_avx512_no_vzeroupper(%rip), %RAX_LP

2
sysdeps/x86_64/multiarch/memcpy_chk.S
View File

@ -30,6 +30,8 @@
ENTRY(__memcpy_chk)
.type __memcpy_chk, @gnu_indirect_function
LOAD_RTLD_GLOBAL_RO_RDX
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 1f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 1f
lea __memcpy_chk_avx512_no_vzeroupper(%rip), %RAX_LP

2
sysdeps/x86_64/multiarch/memmove.S
View File

@ -30,6 +30,8 @@ ENTRY(__libc_memmove)
lea __memmove_erms(%rip), %RAX_LP
HAS_ARCH_FEATURE (Prefer_ERMS)
jnz 2f
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 1f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 1f
lea __memmove_avx512_no_vzeroupper(%rip), %RAX_LP

2
sysdeps/x86_64/multiarch/memmove_chk.S
View File

@ -29,6 +29,8 @@
ENTRY(__memmove_chk)
.type __memmove_chk, @gnu_indirect_function
LOAD_RTLD_GLOBAL_RO_RDX
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 1f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 1f
lea __memmove_chk_avx512_no_vzeroupper(%rip), %RAX_LP

2
sysdeps/x86_64/multiarch/mempcpy.S
View File

@ -32,6 +32,8 @@ ENTRY(__mempcpy)
lea __mempcpy_erms(%rip), %RAX_LP
HAS_ARCH_FEATURE (Prefer_ERMS)
jnz 2f
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 1f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 1f
lea __mempcpy_avx512_no_vzeroupper(%rip), %RAX_LP

2
sysdeps/x86_64/multiarch/mempcpy_chk.S
View File

@ -30,6 +30,8 @@
ENTRY(__mempcpy_chk)
.type __mempcpy_chk, @gnu_indirect_function
LOAD_RTLD_GLOBAL_RO_RDX
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 1f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 1f
lea __mempcpy_chk_avx512_no_vzeroupper(%rip), %RAX_LP

2
sysdeps/x86_64/multiarch/memset.S
View File

@ -41,6 +41,8 @@ ENTRY(memset)
jnz L(AVX512F)
lea __memset_avx2_unaligned(%rip), %RAX_LP
L(AVX512F):
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 2f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 2f
lea __memset_avx512_no_vzeroupper(%rip), %RAX_LP

2
sysdeps/x86_64/multiarch/memset_chk.S
View File

@ -38,6 +38,8 @@ ENTRY(__memset_chk)
jnz L(AVX512F)
lea __memset_chk_avx2_unaligned(%rip), %RAX_LP
L(AVX512F):
HAS_ARCH_FEATURE (Prefer_No_AVX512)
jnz 2f
HAS_ARCH_FEATURE (AVX512F_Usable)
jz 2f
lea __memset_chk_avx512_no_vzeroupper(%rip), %RAX_LP