Alexander Motin f22068d91b amd64: Stop using REP MOVSB for backward memmove()s. ... Enhanced REP MOVSB feature of CPUs starting from Ivy Bridge makes REP MOVSB the fastest way to copy memory in most of cases. However Intel Optimization Reference Manual says: "setting the DF to force REP MOVSB to copy bytes from high towards low addresses will expe- rience significant performance degradation". Measurements on Intel Cascade Lake and Alder Lake, same as on AMD Zen3 show that it can drop throughput to as low as 2.5-3.5GB/s, comparing to ~10-30GB/s of REP MOVSQ or hand-rolled loop, used for non-ERMS CPUs. This patch keeps ERMS use for forward ordered memory copies, but removes it for backward overlapped moves where it does not work. This is just a cosmetic sync with kernel, since libc does not use ERMS at this time. Reviewed by: mjg MFC after: 2 weeks		2022-06-16 14:51:50 -04:00
..
bcmp.S	amd64: bring back asm bcmp, shared with memcmp	2022-03-26 09:10:03 +00:00
bcopy.c	amd64: convert libc bcopy to a C func to avoid future bloat	2018-10-13 21:17:28 +00:00
bzero.c	amd64: convert libc bzero to a C func to avoid future bloat	2018-11-15 20:20:39 +00:00
Makefile.inc	amd64: bring back asm bcmp, shared with memcmp	2022-03-26 09:10:03 +00:00
memcmp.S	amd64: bring back asm bcmp, shared with memcmp	2022-03-26 09:10:03 +00:00
memcpy.S	amd64: import updated kernel memmove to libc	2018-10-13 21:15:47 +00:00
memmove.S	amd64: Stop using REP MOVSB for backward memmove()s.	2022-06-16 14:51:50 -04:00
memset.S	amd64: add a note about simd to libc memset, memmove and memcmp	2021-01-31 16:07:19 +00:00
stpcpy.S
strcat.S
strcmp.S
strcpy.c	libc: further adoption of SPDX licensing ID tags.	2017-11-25 17:12:48 +00:00
strlen.S	amd64: import asm strlen into libc	2021-02-23 00:09:55 +00:00