mirror of
https://sourceware.org/git/glibc.git
synced 2024-12-13 06:40:09 +00:00
4f26956d5b
This patch optimizes the performance of memset for A64FX [1] which implements ARMv8-A SVE and has L1 64KB cache per core and L2 8MB cache per NUMA node. The performance optimization makes use of Scalable Vector Register with several techniques such as loop unrolling, memory access alignment, cache zero fill and prefetch. SVE assembler code for memset is implemented as Vector Length Agnostic code so theoretically it can be run on any SOC which supports ARMv8-A SVE standard. We confirmed that all testcases have been passed by running 'make check' and 'make xcheck' not only on A64FX but also on ThunderX2. And also we confirmed that the SVE 512 bit vector register performance is roughly 4 times better than Advanced SIMD 128 bit register and 8 times better than scalar 64 bit register by running 'make bench'. [1] https://github.com/fujitsu/A64FX Reviewed-by: Wilco Dijkstra <Wilco.Dijkstra@arm.com> Reviewed-by: Szabolcs Nagy <Szabolcs.Nagy@arm.com>
79 lines
3.0 KiB
C
79 lines
3.0 KiB
C
/* Enumerate available IFUNC implementations of a function. AARCH64 version.
|
|
Copyright (C) 2017-2021 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
|
|
<https://www.gnu.org/licenses/>. */
|
|
|
|
#include <assert.h>
|
|
#include <string.h>
|
|
#include <wchar.h>
|
|
#include <ldsodefs.h>
|
|
#include <ifunc-impl-list.h>
|
|
#include <init-arch.h>
|
|
#include <stdio.h>
|
|
|
|
/* Maximum number of IFUNC implementations. */
|
|
#define MAX_IFUNC 7
|
|
|
|
size_t
|
|
__libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
|
|
size_t max)
|
|
{
|
|
assert (max >= MAX_IFUNC);
|
|
|
|
size_t i = 0;
|
|
|
|
INIT_ARCH ();
|
|
|
|
/* Support sysdeps/aarch64/multiarch/memcpy.c, memmove.c and memset.c. */
|
|
IFUNC_IMPL (i, name, memcpy,
|
|
IFUNC_IMPL_ADD (array, i, memcpy, 1, __memcpy_thunderx)
|
|
IFUNC_IMPL_ADD (array, i, memcpy, !bti, __memcpy_thunderx2)
|
|
IFUNC_IMPL_ADD (array, i, memcpy, 1, __memcpy_falkor)
|
|
IFUNC_IMPL_ADD (array, i, memcpy, 1, __memcpy_simd)
|
|
#if HAVE_AARCH64_SVE_ASM
|
|
IFUNC_IMPL_ADD (array, i, memcpy, sve, __memcpy_a64fx)
|
|
#endif
|
|
IFUNC_IMPL_ADD (array, i, memcpy, 1, __memcpy_generic))
|
|
IFUNC_IMPL (i, name, memmove,
|
|
IFUNC_IMPL_ADD (array, i, memmove, 1, __memmove_thunderx)
|
|
IFUNC_IMPL_ADD (array, i, memmove, !bti, __memmove_thunderx2)
|
|
IFUNC_IMPL_ADD (array, i, memmove, 1, __memmove_falkor)
|
|
IFUNC_IMPL_ADD (array, i, memmove, 1, __memmove_simd)
|
|
#if HAVE_AARCH64_SVE_ASM
|
|
IFUNC_IMPL_ADD (array, i, memmove, sve, __memmove_a64fx)
|
|
#endif
|
|
IFUNC_IMPL_ADD (array, i, memmove, 1, __memmove_generic))
|
|
IFUNC_IMPL (i, name, memset,
|
|
/* Enable this on non-falkor processors too so that other cores
|
|
can do a comparative analysis with __memset_generic. */
|
|
IFUNC_IMPL_ADD (array, i, memset, (zva_size == 64), __memset_falkor)
|
|
IFUNC_IMPL_ADD (array, i, memset, (zva_size == 64), __memset_emag)
|
|
IFUNC_IMPL_ADD (array, i, memset, 1, __memset_kunpeng)
|
|
#if HAVE_AARCH64_SVE_ASM
|
|
IFUNC_IMPL_ADD (array, i, memset, sve, __memset_a64fx)
|
|
#endif
|
|
IFUNC_IMPL_ADD (array, i, memset, 1, __memset_generic))
|
|
IFUNC_IMPL (i, name, memchr,
|
|
IFUNC_IMPL_ADD (array, i, memchr, !mte, __memchr_nosimd)
|
|
IFUNC_IMPL_ADD (array, i, memchr, 1, __memchr_generic))
|
|
|
|
IFUNC_IMPL (i, name, strlen,
|
|
IFUNC_IMPL_ADD (array, i, strlen, !mte, __strlen_asimd)
|
|
IFUNC_IMPL_ADD (array, i, strlen, 1, __strlen_mte))
|
|
|
|
return i;
|
|
}
|