> > Any chance on a libc-p4? > Any additional optimisations need to have benchmarks proving that > they're worth the hassle. Every one that we add carries the cost of > additional troubleshooting, etc. > Do you have data that suggests that there's a notable improvement tobe > had? For memcpy I found an 8,4% speed improvement when copying 100MB using all 4 execution pipelines instead of 2 currently used.
|
-- Although when you're in the situation that RMS is telling you that you're being too ideological about freedom, maybe, just maybe, it's true. - Matthew Wilcox |
Attachment:
signature.asc
Description: Ceci est une partie de message =?ISO-8859-1?Q?num=E9riquement?= =?ISO-8859-1?Q?_sign=E9e?=