Messages in this thread | | | Date | Thu, 19 Apr 2012 10:02:28 +1000 | From | NeilBrown <> | Subject | Re: arch/86: AVX RAID5 xor checksumming v1 |
| |
On Wed, 18 Apr 2012 15:58:47 -0700 Jim Kukunas <james.t.kukunas@linux.intel.com> wrote:
> Hi Folks, > > The following patch adds an AVX implementation of the RAID5 xor checksumming > functions. > > This version differs from version 0, in that it: > 0) checks whether assembler supports AVX > 1) replaces __attribute__(aligned(32)) with __aligned(32) > 2) adds do/while(0); to macro blocks > > Based on xor_speed, the AVX implementation appears to be ~32% faster than the > SSE implementation on my i7 2600: > > generic_sse: 15088.000 MB/sec > avx: 19936.000 MB/sec >
Applied, thanks. (though I remove the blank line at the end of the new file - the voices made me do it, I had no choice)
NeilBrown [unhandled content-type:application/pgp-signature] | |