String functions make little difference it seems. The critical ones are
copy_to_user copy_from_user memcpy memset and the copy_page/clear_page
functions (mapped to memset etc on x86)
The intel cpus have fastpaths for rep movs so for everything but the 486
and some pentiums it seems that the 386 code is fastest. On a K6-II I can't
beat rep movs for memory originated reads. Im not sure about cached stuff.
Some people apparently get gains on pentiums using mmx/fpu for memcpy.
The Athlon seems unique in having MMX deliver measurably faster performance
than rep movs
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/