diff options
| author | H.J. Lu <hjl.tools@gmail.com> | 2020-01-25 14:19:40 -0800 |
|---|---|---|
| committer | H.J. Lu <hjl.tools@gmail.com> | 2021-01-04 07:58:57 -0800 |
| commit | 3ec5d83d2a237d39e7fd6ef7a0bc8ac4c171a4a5 (patch) | |
| tree | 6d2205ddecade9c780d12c37a809a3c8d4fc3384 /sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S | |
| parent | cd6274089f7a7603cdaf2a24cef575fa61d3772e (diff) | |
| download | glibc-3ec5d83d2a237d39e7fd6ef7a0bc8ac4c171a4a5.tar.gz glibc-3ec5d83d2a237d39e7fd6ef7a0bc8ac4c171a4a5.tar.xz glibc-3ec5d83d2a237d39e7fd6ef7a0bc8ac4c171a4a5.zip | |
x86-64: Avoid rep movsb with short distance [BZ #27130]
When copying with "rep movsb", if the distance between source and destination is N*4GB + [1..63] with N >= 0, performance may be very slow. This patch updates memmove-vec-unaligned-erms.S for AVX and AVX512 versions with the distance in RCX: cmpl $63, %ecx // Don't use "rep movsb" if ECX <= 63 jbe L(Don't use rep movsb") Use "rep movsb" Benchtests data with bench-memcpy, bench-memcpy-large, bench-memcpy-random and bench-memcpy-walk on Skylake, Ice Lake and Tiger Lake show that its performance impact is within noise range as "rep movsb" is only used for data size >= 4KB.
Diffstat (limited to 'sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S')
| -rw-r--r-- | sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S | 21 |
1 files changed, 21 insertions, 0 deletions
diff --git a/sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S b/sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S index 7d54095f04..0980c95378 100644 --- a/sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S +++ b/sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S @@ -56,6 +56,13 @@ # endif #endif +/* Avoid short distance rep movsb only with non-SSE vector. */ +#ifndef AVOID_SHORT_DISTANCE_REP_MOVSB +# define AVOID_SHORT_DISTANCE_REP_MOVSB (VEC_SIZE > 16) +#else +# define AVOID_SHORT_DISTANCE_REP_MOVSB 0 +#endif + #ifndef PREFETCH # define PREFETCH(addr) prefetcht0 addr #endif @@ -243,7 +250,21 @@ L(movsb): cmpq %r9, %rdi /* Avoid slow backward REP MOVSB. */ jb L(more_8x_vec_backward) +# if AVOID_SHORT_DISTANCE_REP_MOVSB + movq %rdi, %rcx + subq %rsi, %rcx + jmp 2f +# endif 1: +# if AVOID_SHORT_DISTANCE_REP_MOVSB + movq %rsi, %rcx + subq %rdi, %rcx +2: +/* Avoid "rep movsb" if RCX, the distance between source and destination, + is N*4GB + [1..63] with N >= 0. */ + cmpl $63, %ecx + jbe L(more_2x_vec) /* Avoid "rep movsb" if ECX <= 63. */ +# endif mov %RDX_LP, %RCX_LP rep movsb L(nop): |