PATCH: Improve 64bit strcpy functions

[H.J. Lu]

Hi,

This patch has new SSE2/SSSE3 implementations of 64bit strcpy functions.
Tested on Atom, Core 2 and Core i7.  It improves strcpy performance by
up to 4X.

BTW, this patch needs init-arch changes in the 32bit strcpy patch.


H.J.
----
2011-06-22  H.J. Lu  <hongjiu.lu@intel.com>

	* sysdeps/x86_64/multiarch/Makefile (sysdep_routines): Add
	strcpy-ssse3 strncpy-ssse3 stpcpy-ssse3 stpncpy-ssse3
	strcpy-sse2-unaligned strncpy-sse2-unaligned
	stpcpy-sse2-unaligned stpncpy-sse2-unaligned.

	* sysdeps/x86_64/multiarch/stpcpy-sse2-unaligned.S: New.
	* sysdeps/x86_64/multiarch/stpcpy-ssse3.S: Likewise.
	* sysdeps/x86_64/multiarch/stpncpy-sse2-unaligned.S: Likewise.
	* sysdeps/x86_64/multiarch/stpncpy-ssse3.S: Likewise.
	* sysdeps/x86_64/multiarch/strcpy-sse2-unaligned.S: Likewise.
	* sysdeps/x86_64/multiarch/strcpy-ssse3.S: Likewise.
	* sysdeps/x86_64/multiarch/strncpy-sse2-unaligned.S: Likewise.
	* sysdeps/x86_64/multiarch/strncpy-ssse3.S: Likewise.

	* sysdeps/x86_64/multiarch/strcpy.S: Remove strcpy with SSSE3.
	(STRCPY): Support SSE2 and SSSE3 versions.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: libc-strcpy-64bit-1.patch.bz2
Type: application/x-bzip2
Size: 11792 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/libc-alpha/attachments/20110623/c104d386/attachment.bz2>