diff options
| author | Lynne <dev@lynne.ee> | 2023-02-01 02:26:20 +0100 |
|---|---|---|
| committer | Lynne <dev@lynne.ee> | 2023-02-01 04:23:55 +0100 |
| commit | bbe95f7353a972f28a48be8da883549f02c59e4b (patch) | |
| tree | 08841c9da55e7f076f6046d1dbd70f49d74c0ec0 /libavfilter/x86/vf_w3fdif.asm | |
| parent | fc9a3b584da3cf3fc1f00036be2eaf5dff903ccf (diff) | |
| download | ffmpeg-bbe95f7353a972f28a48be8da883549f02c59e4b.tar.gz | |
x86: replace explicit REP_RETs with RETs
From x86inc:
> On AMD cpus <=K10, an ordinary ret is slow if it immediately follows either
> a branch or a branch target. So switch to a 2-byte form of ret in that case.
> We can automatically detect "follows a branch", but not a branch target.
> (SSSE3 is a sufficient condition to know that your cpu doesn't have this problem.)
x86inc can automatically determine whether to use REP_RET rather than
REP in most of these cases, so impact is minimal. Additionally, a few
REP_RETs were used unnecessary, despite the return being nowhere near a
branch.
The only CPUs affected were AMD K10s, made between 2007 and 2011, 16
years ago and 12 years ago, respectively.
In the future, everyone involved with x86inc should consider dropping
REP_RETs altogether.
Diffstat (limited to 'libavfilter/x86/vf_w3fdif.asm')
| -rw-r--r-- | libavfilter/x86/vf_w3fdif.asm | 10 |
1 files changed, 5 insertions, 5 deletions
diff --git a/libavfilter/x86/vf_w3fdif.asm b/libavfilter/x86/vf_w3fdif.asm index 52628c38d7..3010469f97 100644 --- a/libavfilter/x86/vf_w3fdif.asm +++ b/libavfilter/x86/vf_w3fdif.asm @@ -38,7 +38,7 @@ cglobal w3fdif_scale, 3, 3, 2, 0, out_pixel, work_pixel, linesize add work_pixelq, mmsize*2 sub linesized, mmsize/2 jg .loop -REP_RET +RET cglobal w3fdif_simple_low, 4, 5, 6, 0, work_line, in_lines_cur0, coef, linesize, offset movd m1, [coefq] @@ -63,7 +63,7 @@ cglobal w3fdif_simple_low, 4, 5, 6, 0, work_line, in_lines_cur0, coef, linesize, add offsetq, mmsize/2 sub linesized, mmsize/2 jg .loop -REP_RET +RET cglobal w3fdif_complex_low, 4, 7, 8, 0, work_line, in_lines_cur0, coef, linesize movq m0, [coefq] @@ -99,7 +99,7 @@ cglobal w3fdif_complex_low, 4, 7, 8, 0, work_line, in_lines_cur0, coef, linesize add offsetq, mmsize/2 sub linesized, mmsize/2 jg .loop -REP_RET +RET %if ARCH_X86_64 cglobal w3fdif_simple_high, 5, 9, 8, 0, work_line, in_lines_cur0, in_lines_adj0, coef, linesize @@ -179,7 +179,7 @@ cglobal w3fdif_simple_high, 4, 7, 8, 0, work_line, in_lines_cur0, in_lines_adj0, add offsetq, mmsize/2 sub linesized, mmsize/2 jg .loop -REP_RET +RET %if ARCH_X86_64 @@ -254,6 +254,6 @@ cglobal w3fdif_complex_high, 5, 13, 10, 0, work_line, in_lines_cur0, in_lines_ad add offsetq, mmsize/2 sub linesized, mmsize/2 jg .loop -REP_RET +RET %endif |