aboutsummaryrefslogtreecommitdiffstats
path: root/libavcodec/mips
diff options
context:
space:
mode:
authorChristophe GISQUET <christophe.gisquet@gmail.com>2012-02-23 19:48:58 +0100
committerRonald S. Bultje <rsbultje@gmail.com>2012-02-23 15:50:06 -0800
commit34454c761f01275d4adaf40df6d70a59011c4a6c (patch)
treea25a23c028ddee97c1195567f855ce064bdbe916 /libavcodec/mips
parent2e74a5abc2fda6cfbc86589852d6194d502332cb (diff)
downloadffmpeg-34454c761f01275d4adaf40df6d70a59011c4a6c.tar.gz
SBR DSP x86: implement SSE sbr_sum_square_sse
The 32bits targets have been compiled with -mfpmath=sse for proper reference. sbr_sum_square C /32bits: 82c (unrolled)/102c C /64bits: 69c (unrolled)/82c SSE/32bits: 42c SSE/64bits: 31c Use of SSE4.1 dpps to perform the final sum is slower. Not unrolling to perform 8 operations in a loop yields 10 more cycles. Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
Diffstat (limited to 'libavcodec/mips')
0 files changed, 0 insertions, 0 deletions