This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] x86-64: Optimize strrchr/wcsrchr with AVX2
- From: "H.J. Lu" <hjl dot tools at gmail dot com>
- To: Adhemerval Zanella <adhemerval dot zanella at linaro dot org>
- Cc: GNU C Library <libc-alpha at sourceware dot org>
- Date: Fri, 2 Jun 2017 12:52:41 -0700
- Subject: Re: [PATCH] x86-64: Optimize strrchr/wcsrchr with AVX2
- Authentication-results: sourceware.org; auth=none
- References: <20170601181401.GD28627@lucon.org> <34399498-351f-2d58-b7b2-8119dcd0f142@linaro.org>
On Thu, Jun 1, 2017 at 11:42 AM, Adhemerval Zanella
<adhemerval.zanella@linaro.org> wrote:
>
>
> On 01/06/2017 15:14, H.J. Lu wrote:
>> Optimize strrchr/wcsrchr with AVX2 to check 32 bytes with vector
>> instructions. It is as fast as SSE2 version for small data sizes
>> and up to 1X faster for large data sizes on Haswell. Select AVX2
>> version on AVX2 machines where vzeroupper is preferred and AVX
>> unaligned load is fast.
>>
>> Any comments?
>>
>> H.J.
>> --
>> * sysdeps/x86_64/multiarch/Makefile (sysdep_routines): Add
>> strrchr-avx2 and wcsrchr-avx2.
>> * sysdeps/x86_64/multiarch/ifunc-impl-list.c
>> (__libc_ifunc_impl_list): Add tests for __strrchr_avx2,
>> __strrchr_sse2, __wcsrchr_avx2 and __wcsrchr_sse2.
>> * sysdeps/x86_64/multiarch/strrchr-avx2.S: New file.
>> * sysdeps/x86_64/multiarch/strrchr.S: Likewise.
>
> I think this could be an opportunity to avoid adding more assembly ifunc
> written in assembly and use C implementation instead.
Good idea. Here is the updated patch.
Thanks.
--
H.J.
From 1b2001d189cfa3327761ab7b76c47a48e3782a04 Mon Sep 17 00:00:00 2001
From: "H.J. Lu" <hjl.tools@gmail.com>
Date: Fri, 26 May 2017 12:21:55 -0700
Subject: [PATCH 8/8] x86-64: Optimize strrchr/wcsrchr with AVX2
Optimize strrchr/wcsrchr with AVX2 to check 32 bytes with vector
instructions. It is as fast as SSE2 version for small data sizes
and up to 1X faster for large data sizes on Haswell. Select AVX2
version on AVX2 machines where vzeroupper is preferred and AVX
unaligned load is fast.
* sysdeps/x86_64/multiarch/Makefile (sysdep_routines): Add
strrchr-sse2, strrchr-avx2, wcsrchr-sse2 and wcsrchr-avx2.
* sysdeps/x86_64/multiarch/ifunc-impl-list.c
(__libc_ifunc_impl_list): Add tests for __strrchr_avx2,
__strrchr_sse2, __wcsrchr_avx2 and __wcsrchr_sse2.
* sysdeps/x86_64/multiarch/strrchr-avx2.S: New file.
* sysdeps/x86_64/multiarch/strrchr-sse2.S: Likewise.
* sysdeps/x86_64/multiarch/strrchr.c: Likewise.
* sysdeps/x86_64/multiarch/wcsrchr-avx2.S: Likewise.
* sysdeps/x86_64/multiarch/wcsrchr-sse2.S: Likewise.
* sysdeps/x86_64/multiarch/wcsrchr.c: Likewise.
---
sysdeps/x86_64/multiarch/Makefile | 2 +
sysdeps/x86_64/multiarch/ifunc-impl-list.c | 14 ++
sysdeps/x86_64/multiarch/strrchr-avx2.S | 235 +++++++++++++++++++++++++++++
sysdeps/x86_64/multiarch/strrchr-sse2.S | 35 +++++
sysdeps/x86_64/multiarch/strrchr.c | 30 ++++
sysdeps/x86_64/multiarch/wcsrchr-avx2.S | 3 +
sysdeps/x86_64/multiarch/wcsrchr-sse2.S | 23 +++
sysdeps/x86_64/multiarch/wcsrchr.c | 29 ++++
8 files changed, 371 insertions(+)
create mode 100644 sysdeps/x86_64/multiarch/strrchr-avx2.S
create mode 100644 sysdeps/x86_64/multiarch/strrchr-sse2.S
create mode 100644 sysdeps/x86_64/multiarch/strrchr.c
create mode 100644 sysdeps/x86_64/multiarch/wcsrchr-avx2.S
create mode 100644 sysdeps/x86_64/multiarch/wcsrchr-sse2.S
create mode 100644 sysdeps/x86_64/multiarch/wcsrchr.c
diff --git a/sysdeps/x86_64/multiarch/Makefile b/sysdeps/x86_64/multiarch/Makefile
index 60c586c..6baf2db 100644
--- a/sysdeps/x86_64/multiarch/Makefile
+++ b/sysdeps/x86_64/multiarch/Makefile
@@ -15,6 +15,7 @@ sysdep_routines += strncat-c stpncpy-c strncpy-c strcmp-ssse3 \
memmove-ssse3-back \
memmove-avx512-no-vzeroupper strcasecmp_l-ssse3 \
strchr-sse2 strchrnul-sse2 strchr-avx2 strchrnul-avx2 \
+ strrchr-sse2 strrchr-avx2 \
strlen-sse2 strnlen-sse2 strlen-avx2 strnlen-avx2 \
strncase_l-ssse3 strcat-ssse3 strncat-ssse3\
strcpy-ssse3 strncpy-ssse3 stpcpy-ssse3 stpncpy-ssse3 \
@@ -40,6 +41,7 @@ sysdep_routines += wmemcmp-sse4 wmemcmp-ssse3 wmemcmp-c \
wmemcmp-avx2-movbe \
wcscpy-ssse3 wcscpy-c \
wcschr-sse2 wcschr-avx2 \
+ wcsrchr-sse2 wcsrchr-avx2 \
wcslen-sse2 wcsnlen-sse2 wcslen-avx2 wcsnlen-avx2
endif
diff --git a/sysdeps/x86_64/multiarch/ifunc-impl-list.c b/sysdeps/x86_64/multiarch/ifunc-impl-list.c
index 86bd4ab..c2578cb 100644
--- a/sysdeps/x86_64/multiarch/ifunc-impl-list.c
+++ b/sysdeps/x86_64/multiarch/ifunc-impl-list.c
@@ -251,6 +251,13 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
__strchrnul_avx2)
IFUNC_IMPL_ADD (array, i, strchrnul, 1, __strchrnul_sse2))
+ /* Support sysdeps/x86_64/multiarch/strrchr.S. */
+ IFUNC_IMPL (i, name, strrchr,
+ IFUNC_IMPL_ADD (array, i, strrchr,
+ HAS_ARCH_FEATURE (AVX2_Usable),
+ __strrchr_avx2)
+ IFUNC_IMPL_ADD (array, i, strrchr, 1, __strrchr_sse2))
+
/* Support sysdeps/x86_64/multiarch/strcmp.S. */
IFUNC_IMPL (i, name, strcmp,
IFUNC_IMPL_ADD (array, i, strcmp, HAS_CPU_FEATURE (SSE4_2),
@@ -342,6 +349,13 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
__wcschr_avx2)
IFUNC_IMPL_ADD (array, i, wcschr, 1, __wcschr_sse2))
+ /* Support sysdeps/x86_64/multiarch/wcsrchr.S. */
+ IFUNC_IMPL (i, name, wcsrchr,
+ IFUNC_IMPL_ADD (array, i, wcsrchr,
+ HAS_ARCH_FEATURE (AVX2_Usable),
+ __wcsrchr_avx2)
+ IFUNC_IMPL_ADD (array, i, wcsrchr, 1, __wcsrchr_sse2))
+
/* Support sysdeps/x86_64/multiarch/wcscpy.S. */
IFUNC_IMPL (i, name, wcscpy,
IFUNC_IMPL_ADD (array, i, wcscpy, HAS_CPU_FEATURE (SSSE3),
diff --git a/sysdeps/x86_64/multiarch/strrchr-avx2.S b/sysdeps/x86_64/multiarch/strrchr-avx2.S
new file mode 100644
index 0000000..36ef660
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/strrchr-avx2.S
@@ -0,0 +1,235 @@
+/* strrchr/wcsrchr optimized with AVX2.
+ Copyright (C) 2017 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#if IS_IN (libc)
+
+# include <sysdep.h>
+
+# ifndef STRRCHR
+# define STRRCHR __strrchr_avx2
+# endif
+
+# ifdef USE_AS_WCSRCHR
+# define VPBROADCAST vpbroadcastd
+# define VPCMPEQ vpcmpeqd
+# else
+# define VPBROADCAST vpbroadcastb
+# define VPCMPEQ vpcmpeqb
+# endif
+
+# ifndef VZEROUPPER
+# define VZEROUPPER vzeroupper
+# endif
+
+# define VEC_SIZE 32
+
+ .section .text.avx,"ax",@progbits
+ENTRY (STRRCHR)
+ movd %esi, %xmm4
+ movl %edi, %ecx
+ /* Broadcast CHAR to YMM4. */
+ VPBROADCAST %xmm4, %ymm4
+ vpxor %ymm0, %ymm0, %ymm0
+
+ /* Check if we may cross page boundary with one vector load. */
+ andl $(2 * VEC_SIZE - 1), %ecx
+ cmpl $VEC_SIZE, %ecx
+ ja L(cros_page_boundary)
+
+ vmovdqu (%rdi), %ymm1
+ VPCMPEQ %ymm1, %ymm0, %ymm2
+ VPCMPEQ %ymm1, %ymm4, %ymm3
+ vpmovmskb %ymm2, %ecx
+ vpmovmskb %ymm3, %eax
+ addq $VEC_SIZE, %rdi
+
+ testl %eax, %eax
+ jnz L(first_vec)
+
+ testl %ecx, %ecx
+ jnz L(return_null)
+
+ andq $-VEC_SIZE, %rdi
+ xorl %edx, %edx
+ jmp L(aligned_loop)
+
+ .p2align 4
+L(first_vec):
+ /* Check if there is a nul CHAR. */
+ testl %ecx, %ecx
+ jnz L(char_and_nul_in_first_vec)
+
+ /* Remember the match and keep searching. */
+ movl %eax, %edx
+ movq %rdi, %rsi
+ andq $-VEC_SIZE, %rdi
+ jmp L(aligned_loop)
+
+ .p2align 4
+L(cros_page_boundary):
+ andl $(VEC_SIZE - 1), %ecx
+ andq $-VEC_SIZE, %rdi
+ vmovdqa (%rdi), %ymm1
+ VPCMPEQ %ymm1, %ymm0, %ymm2
+ VPCMPEQ %ymm1, %ymm4, %ymm3
+ vpmovmskb %ymm2, %edx
+ vpmovmskb %ymm3, %eax
+ shrl %cl, %edx
+ shrl %cl, %eax
+ addq $VEC_SIZE, %rdi
+
+ /* Check if there is a CHAR. */
+ testl %eax, %eax
+ jnz L(found_char)
+
+ testl %edx, %edx
+ jnz L(return_null)
+
+ jmp L(aligned_loop)
+
+ .p2align 4
+L(found_char):
+ testl %edx, %edx
+ jnz L(char_and_nul)
+
+ /* Remember the match and keep searching. */
+ movl %eax, %edx
+ leaq (%rdi, %rcx), %rsi
+
+ .p2align 4
+L(aligned_loop):
+ vmovdqa (%rdi), %ymm1
+ VPCMPEQ %ymm1, %ymm0, %ymm2
+ addq $VEC_SIZE, %rdi
+ VPCMPEQ %ymm1, %ymm4, %ymm3
+ vpmovmskb %ymm2, %ecx
+ vpmovmskb %ymm3, %eax
+ orl %eax, %ecx
+ jnz L(char_nor_null)
+
+ vmovdqa (%rdi), %ymm1
+ VPCMPEQ %ymm1, %ymm0, %ymm2
+ add $VEC_SIZE, %rdi
+ VPCMPEQ %ymm1, %ymm4, %ymm3
+ vpmovmskb %ymm2, %ecx
+ vpmovmskb %ymm3, %eax
+ orl %eax, %ecx
+ jnz L(char_nor_null)
+
+ vmovdqa (%rdi), %ymm1
+ VPCMPEQ %ymm1, %ymm0, %ymm2
+ addq $VEC_SIZE, %rdi
+ VPCMPEQ %ymm1, %ymm4, %ymm3
+ vpmovmskb %ymm2, %ecx
+ vpmovmskb %ymm3, %eax
+ orl %eax, %ecx
+ jnz L(char_nor_null)
+
+ vmovdqa (%rdi), %ymm1
+ VPCMPEQ %ymm1, %ymm0, %ymm2
+ addq $VEC_SIZE, %rdi
+ VPCMPEQ %ymm1, %ymm4, %ymm3
+ vpmovmskb %ymm2, %ecx
+ vpmovmskb %ymm3, %eax
+ orl %eax, %ecx
+ jz L(aligned_loop)
+
+ .p2align 4
+L(char_nor_null):
+ /* Find a CHAR or a nul CHAR in a loop. */
+ testl %eax, %eax
+ jnz L(match)
+L(return_value):
+ testl %edx, %edx
+ jz L(return_null)
+ movl %edx, %eax
+ movq %rsi, %rdi
+
+# ifdef USE_AS_WCSRCHR
+ /* Keep the first bit for each matching CHAR for bsr. */
+ andl $0x11111111, %eax
+# endif
+ bsrl %eax, %eax
+ leaq -VEC_SIZE(%rdi, %rax), %rax
+ VZEROUPPER
+ ret
+
+ .p2align 4
+L(match):
+ /* Find a CHAR. Check if there is a nul CHAR. */
+ vpmovmskb %ymm2, %ecx
+ testl %ecx, %ecx
+ jnz L(find_nul)
+
+ /* Remember the match and keep searching. */
+ movl %eax, %edx
+ movq %rdi, %rsi
+ jmp L(aligned_loop)
+
+ .p2align 4
+L(find_nul):
+# ifdef USE_AS_WCSRCHR
+ /* Keep the first bit for each matching CHAR for bsr. */
+ andl $0x11111111, %ecx
+ andl $0x11111111, %eax
+# endif
+ /* Mask out any matching bits after the nul CHAR. */
+ movl %ecx, %r8d
+ subl $1, %r8d
+ xorl %ecx, %r8d
+ andl %r8d, %eax
+ testl %eax, %eax
+ /* If there is no CHAR here, return the remembered one. */
+ jz L(return_value)
+ bsrl %eax, %eax
+ leaq -VEC_SIZE(%rdi, %rax), %rax
+ VZEROUPPER
+ ret
+
+ .p2align 4
+L(char_and_nul):
+ /* Find both a CHAR and a nul CHAR. */
+ addq %rcx, %rdi
+ movl %edx, %ecx
+L(char_and_nul_in_first_vec):
+# ifdef USE_AS_WCSRCHR
+ /* Keep the first bit for each matching CHAR for bsr. */
+ andl $0x11111111, %ecx
+ andl $0x11111111, %eax
+# endif
+ /* Mask out any matching bits after the nul CHAR. */
+ movl %ecx, %r8d
+ subl $1, %r8d
+ xorl %ecx, %r8d
+ andl %r8d, %eax
+ testl %eax, %eax
+ /* Return null pointer if the nul CHAR comes first. */
+ jz L(return_null)
+ bsrl %eax, %eax
+ leaq -VEC_SIZE(%rdi, %rax), %rax
+ VZEROUPPER
+ ret
+
+ .p2align 4
+L(return_null):
+ xorl %eax, %eax
+ VZEROUPPER
+ ret
+
+END (STRRCHR)
+#endif
diff --git a/sysdeps/x86_64/multiarch/strrchr-sse2.S b/sysdeps/x86_64/multiarch/strrchr-sse2.S
new file mode 100644
index 0000000..9e8dc2a
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/strrchr-sse2.S
@@ -0,0 +1,35 @@
+/* strrchr optimized with SSE2.
+ Copyright (C) 2017 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#if IS_IN (libc)
+# define strrchr __strrchr_sse2
+
+# ifdef SHARED
+# undef libc_hidden_builtin_def
+/* It doesn't make sense to send libc-internal strrchr calls through a PLT.
+ The speedup we get from using SSE4.2 instruction is likely eaten away
+ by the indirect call in the PLT. */
+# define libc_hidden_builtin_def(name) \
+ .globl __GI_strrchr; __GI_strrchr = __strrchr_sse2
+# endif
+
+# undef weak_alias
+# define weak_alias(strrchr, rindex)
+#endif
+
+#include "../strrchr.S"
diff --git a/sysdeps/x86_64/multiarch/strrchr.c b/sysdeps/x86_64/multiarch/strrchr.c
new file mode 100644
index 0000000..fbb9c84
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/strrchr.c
@@ -0,0 +1,30 @@
+/* Multiple versions of strrchr.
+ Copyright (C) 2017 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+/* Define multiple versions only for the definition in libc. */
+#if IS_IN (libc)
+# define strrchr __redirect_strrchr
+# include <string.h>
+# undef strrchr
+
+# define SYMBOL_NAME strrchr
+# include "ifunc-sse2-avx2.h"
+
+libc_ifunc_redirected (__redirect_strrchr, strrchr, IFUNC_SELECTOR ());
+weak_alias (strrchr, rindex);
+#endif
diff --git a/sysdeps/x86_64/multiarch/wcsrchr-avx2.S b/sysdeps/x86_64/multiarch/wcsrchr-avx2.S
new file mode 100644
index 0000000..cf8a239
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/wcsrchr-avx2.S
@@ -0,0 +1,3 @@
+#define STRRCHR __wcsrchr_avx2
+#define USE_AS_WCSRCHR 1
+#include "strrchr-avx2.S"
diff --git a/sysdeps/x86_64/multiarch/wcsrchr-sse2.S b/sysdeps/x86_64/multiarch/wcsrchr-sse2.S
new file mode 100644
index 0000000..0ac1b13
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/wcsrchr-sse2.S
@@ -0,0 +1,23 @@
+/* wcsrchr optimized with SSE2.
+ Copyright (C) 2017 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#if IS_IN (libc)
+# define wcsrchr __wcsrchr_sse2
+#endif
+
+#include "../wcsrchr.S"
diff --git a/sysdeps/x86_64/multiarch/wcsrchr.c b/sysdeps/x86_64/multiarch/wcsrchr.c
new file mode 100644
index 0000000..0413bd2
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/wcsrchr.c
@@ -0,0 +1,29 @@
+/* Multiple versions of wcsrchr.
+ Copyright (C) 2017 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+/* Define multiple versions only for the definition in libc. */
+#if IS_IN (libc)
+# define wcsrchr __redirect_wcsrchr
+# include <wchar.h>
+# undef wcsrchr
+
+# define SYMBOL_NAME wcsrchr
+# include "ifunc-sse2-avx2.h"
+
+libc_ifunc_redirected (__redirect_wcsrchr, wcsrchr, IFUNC_SELECTOR ());
+#endif
--
2.9.4