diff options
| author |  mtklein <mtklein@chromium.org> | 2016-06-07 09:35:27 -0700 |
|---|---|---|
| committer |  Commit bot <commit-bot@chromium.org> | 2016-06-07 09:35:28 -0700 |
| commit | 12dfaaa53c23f3d03050bde8f64136ac1f44164a (patch) | |
| tree | 63cfa96123575974f0560f785b3bc63367e15e63 /src/opts/SkNx_sse.h | |
| parent | d62e28b19a23b913c549b7891ecf79e779577181 (diff) | |
Move immintrin/arm_neon includes to where they are used.
On my Mac (so, immintrin), this improves compile time, both wall and cpu,
by about 16%. To test I ran this on an SSD with files hot in their caches:
$ env CC=/usr/bin/clang CXX=/usr/bin/clang++ ./gyp_skia && \
ninja -C out/Release -t clean && \
time ninja -C out/Release
Before: 159 wall / 3367 cpu
159 wall / 3368 cpu
After: 137 wall / 2860 cpu
136 wall / 2863 cpu
I also tried further refining immintrin down to emmintrin / tmmintrin / smmintrin etc.
That made no signficant difference, so I've kept immintrin for its simplicity.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2045633002
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
TBR=reed@google.com
No public API changes.
Review-Url: https://codereview.chromium.org/2045633002
Diffstat (limited to 'src/opts/SkNx_sse.h')
| -rw-r--r-- | src/opts/SkNx_sse.h | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/src/opts/SkNx_sse.h b/src/opts/SkNx_sse.h index 1fc235d99b..0b22a5a3c6 100644 --- a/src/opts/SkNx_sse.h +++ b/src/opts/SkNx_sse.h @@ -9,6 +9,7 @@ #define SkNx_sse_DEFINED #include "SkCpu.h" +#include <immintrin.h> // This file may assume <= SSE2, but must check SK_CPU_SSE_LEVEL for anything more recent. // If you do, make sure this is in a static inline function... anywhere else risks violating ODR. |