This is a follow-up to https://codereview.chromium.org/1290423007/,
with a couple small changes:
- turn on AVX and AVX2 for Windows using /arch ('EnabledEnhancedInstructionSet')
- reformat and de-conditionalize where possible / irrelevant
Picked up this while poking around in libvpx's Chrome GYPs.
And yes, AVX = 3, AVX2 = 5. Don't even ask what 4 means...
BUG=skia:4117
Review URL: https://codereview.chromium.org/1309253002
CL (1 of 3) adds empty lists in our .gypi,
and builds the files in those empty lists with the appropriate flags.
CL (2 of 3) will have Chrome's GYP and GN files read these lists,
and build them with the appropriate flags.
CL (3 of 3) will add runtime detection and stub files to the lists
with empty Init_sse42(), Init_avx(), Init_avx2() methods.
After that, we should be able to use SSE 4.2, AVX, and AVX2 if desired.
Some motivation:
- SSE 4.2 adds some sweet string-oriented methods that
can help us write fast high quality 32-bit hashes.
- AVX is SSE doubled, e.g. 8 floats or two SkPMFloat at a time.
- AVX2 is SSE2 doubled, e.g. 8 pixels at a time.
BUG=skia:4117
Review URL: https://codereview.chromium.org/1290423007
Pretty vanilla translation. I cleaned up who calls whom a little.
Used to be utils -> opts -> utils, now it's just utils -> opts.
I may follow up with a pass over the NEON code for readability
and to clean up dead code.
This turns on NEON A8->R11EAC conversion for ARMv8.
Unit tests which now hit the NEON code still pass.
I can't find any related bench.
BUG=skia:4117
Review URL: https://codereview.chromium.org/1273103002
I'll be moving headers from src/core to include/private, so this guarantees
that anyone who was finding them via -Isrc/core can now find them via
-Iinclude/private.
This is purely mechanical, mostly to preserve my sanity, so it's likely
(harmless) overkill.
Chromium's GYP and GN builds already set -Iinclude/private for Skia builds.
BUG=skia:4126
Review URL: https://codereview.chromium.org/1265443002
This doesn't really do anything yet. It's just the CPU detection code, skeleton new .cpp files, and a few little .gyp tweaks.
BUG=skia:4117
Review URL: https://codereview.chromium.org/1255193002
(We never set arm_version = 7 for iOS...)
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Mac10.7-Clang-Arm7-Debug-iOS-Trybot
BUG=skia:
Review URL: https://codereview.chromium.org/1092753002
Step 1 of a zillion in the quest for NEON on iOS,
and step 1 of a different zillion in the Great Assembly Purge.
ios, arm, arm64, arm_v7, arm_v7_neon all build.
BUG=skia:
Review URL: https://codereview.chromium.org/1072063002
When checking the skia_arch_type for "x86", instead of doing an
== compare, check if "x86" in skia_arch_type, so it will cover
both x86 and x86_64.
Except when we specifically want x86.
Set skia_arch_width based on "64" in skia_arch_type. No need to specify
in scripts.
In gyp_to_android.py, create a separate var_dict for x86_64.
BUG=skia:3419
Review URL: https://codereview.chromium.org/916113002
Reason for revert:
Going to punt on 16-bit float support for now. Can't figure out ARM 64.
Original issue's description:
> GYP groudwork for half-float opts support.
>
> This sets us up two new opts targets with the immediate goal of adding half-float (SkHalf.h) opts:
> - opts_neon_fp16: uses hardware support on most ARM chips with NEON to do 4 conversions at a time;
> - opts_avx: uses hardware support on Intel chips with AVX to do 8 conversions at a time.
>
> opts_avx will be a handy thing to have around later too, especially if we want to work with floats.
>
> This doesn't actually add any new source files to these libraries yet, so they're no-ops for now.
> I'll need to write a parallel change to Chrome's GN and GYPs before we can start adding sources.
>
> This also rolls GYP up to head, to get suppport for EnableEnhancedInstructionSet: '3' on Windows,
> which is how we turn on AVX there. There's no Mac-specific flag, so we use OTHER_CPLUSPLUSFLAGS.
>
> BUG=skia:
>
> TBR=reed@google.com
>
> Committed: https://skia.googlesource.com/skia/+/46b80833394d7919cadf2abf2b93802141dd21c5TBR=reed@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/912223002
This sets us up two new opts targets with the immediate goal of adding half-float (SkHalf.h) opts:
- opts_neon_fp16: uses hardware support on most ARM chips with NEON to do 4 conversions at a time;
- opts_avx: uses hardware support on Intel chips with AVX to do 8 conversions at a time.
opts_avx will be a handy thing to have around later too, especially if we want to work with floats.
This doesn't actually add any new source files to these libraries yet, so they're no-ops for now.
I'll need to write a parallel change to Chrome's GN and GYPs before we can start adding sources.
This also rolls GYP up to head, to get suppport for EnableEnhancedInstructionSet: '3' on Windows,
which is how we turn on AVX there. There's no Mac-specific flag, so we use OTHER_CPLUSPLUSFLAGS.
BUG=skia:
TBR=reed@google.com
Review URL: https://codereview.chromium.org/915693002
Reason for revert:
This kills Mac 10.6 bots.
FAILED: c++ -MMD -MF obj/src/opts/opts_sse4.SkBlitRow_opts_SSE4.o.d -DSK_INTERNAL -DSK_GAMMA_SRGB -DSK_GAMMA_APPLY_TO_A8 -DSK_SCALAR_TO_FLOAT_EXCLUDED -DSK_ALLOW_STATIC_GLOBAL_INITIALIZERS=1 -DSK_SUPPORT_GPU=1 -DSK_SUPPORT_OPENCL=0 -DSK_FORCE_DISTANCE_FIELD_TEXT=0 -DSK_BUILD_FOR_MAC -DSK_CRASH_HANDLER -DSK_DEVELOPER=1 -I../../src/core -I../../src/utils -I../../include/c -I../../include/config -I../../include/core -I../../include/pathops -I../../include/pipe -I../../include/utils/mac -I../../include/effects -O0 -gdwarf-2 -mmacosx-version-min=10.6 -arch x86_64 -mssse3 -Wall -Wextra -Winit-self -Wpointer-arith -Wsign-compare -Wno-unused-parameter -Wno-invalid-offsetof -msse4.1 -c ../../src/opts/SkBlitRow_opts_SSE4.cpp -o obj/src/opts/opts_sse4.SkBlitRow_opts_SSE4.o
../../src/opts/SkBlitRow_opts_SSE4.cpp:15:27: warning: x86intrin.h: No such file or directory
../../src/opts/SkBlitRow_opts_SSE4.cpp: In function 'void S32A_Opaque_BlitRow32_SSE4(SkPMColor*, const SkPMColor*, int, U8CPU)':
../../src/opts/SkBlitRow_opts_SSE4.cpp:40: error: '_mm_testz_si128' was not declared in this scope
../../src/opts/SkBlitRow_opts_SSE4.cpp:45: error: '_mm_testc_si128' was not declared in this scope
Original issue's description:
> SSE4 opaque blend using intrinsics instead of assembly.
>
> Since we had such a hard time with the assembly versions of this blit (to the
> point that we have them completely disabled everywhere), I thought I'd take
> a shot at writing a version of the blit using intrinsics.
>
> The key feature of SSE4 we're exploiting is that we can use ptest (_mm_test*)
> to skip the blend when the 16 src pixels we consider each loop are all opaque
> or all transparent. _mm_shuffle_epi8 from SSSE3 also lends a hand to extract
> all those alphas.
>
> It's worth looking to see if we can backport this type of logic to SSE2 using
> _mm_movemask_epi8, or up to 32 pixels at a time using AVX.
>
> My local performance testing doesn't show this to be an unambiguous win
> (there are probably microbenchmarks and SKPs where we'd be better off just
> powering through the blend rather than looking at alphas), but the potential
> does seem tantalizing enough to let skiaperf vet it on the bots. (< 1.0x is a win.)
>
> DM says it draws pixel perfect compare to the old code.
>
> Microbenchmarks:
> bitmap_RGBA_8888_A_source_stripes_two 14us -> 14.4us 1.03x
> bitmap_RGBA_8888_A_source_stripes_three 14.3us -> 14.5us 1.01x
> bitmap_RGBA_8888_scale_bilerp 61.9us -> 62.2us 1.01x
> bitmap_RGBA_8888_update_volatile_scale_rotate_bilerp 102us -> 101us 0.99x
> bitmap_RGBA_8888_scale_rotate_bilerp 103us -> 101us 0.99x
> bitmap_RGBA_8888_scale 18.4us -> 18.2us 0.99x
> bitmap_RGBA_8888_A_scale_rotate_bicubic 71us -> 70us 0.99x
> bitmap_RGBA_8888_update_scale_rotate_bilerp 103us -> 101us 0.99x
> bitmap_RGBA_8888_A_scale_rotate_bilerp 112us -> 109us 0.98x
> bitmap_RGBA_8888_update_volatile 5.72us -> 5.58us 0.98x
> bitmap_RGBA_8888 5.73us -> 5.58us 0.97x
> bitmap_RGBA_8888_update 5.78us -> 5.6us 0.97x
> bitmap_RGBA_8888_A_scale_bilerp 70.7us -> 68us 0.96x
> bitmap_RGBA_8888_A_scale_bicubic 23.7us -> 21.8us 0.92x
> bitmap_RGBA_8888_A 13.9us -> 10.9us 0.78x
> bitmap_RGBA_8888_A_source_opaque 14us -> 6.29us 0.45x
> bitmap_RGBA_8888_A_source_transparent 14us -> 3.65us 0.26x
>
> Running over our ~70 SKP web page captures, this looks like we spend 0.7x
> the time in S32A_Opaque_BlitRow compared to the SSE2 version, which should
> be a decent predictor of real-world impact.
>
> BUG=chromium:399842
>
> Committed: https://skia.googlesource.com/skia/+/04bc91b972417038fecfa87c484771eac2b9b785TBR=henrik.smiding@intel.com,mtklein@google.com,herb@google.com,reed@google.com,thakis@chromium.org,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:399842
Review URL: https://codereview.chromium.org/874033004
Since we had such a hard time with the assembly versions of this blit (to the
point that we have them completely disabled everywhere), I thought I'd take
a shot at writing a version of the blit using intrinsics.
The key feature of SSE4 we're exploiting is that we can use ptest (_mm_test*)
to skip the blend when the 16 src pixels we consider each loop are all opaque
or all transparent. _mm_shuffle_epi8 from SSSE3 also lends a hand to extract
all those alphas.
It's worth looking to see if we can backport this type of logic to SSE2 using
_mm_movemask_epi8, or up to 32 pixels at a time using AVX.
My local performance testing doesn't show this to be an unambiguous win
(there are probably microbenchmarks and SKPs where we'd be better off just
powering through the blend rather than looking at alphas), but the potential
does seem tantalizing enough to let skiaperf vet it on the bots. (< 1.0x is a win.)
DM says it draws pixel perfect compare to the old code.
Microbenchmarks:
bitmap_RGBA_8888_A_source_stripes_two 14us -> 14.4us 1.03x
bitmap_RGBA_8888_A_source_stripes_three 14.3us -> 14.5us 1.01x
bitmap_RGBA_8888_scale_bilerp 61.9us -> 62.2us 1.01x
bitmap_RGBA_8888_update_volatile_scale_rotate_bilerp 102us -> 101us 0.99x
bitmap_RGBA_8888_scale_rotate_bilerp 103us -> 101us 0.99x
bitmap_RGBA_8888_scale 18.4us -> 18.2us 0.99x
bitmap_RGBA_8888_A_scale_rotate_bicubic 71us -> 70us 0.99x
bitmap_RGBA_8888_update_scale_rotate_bilerp 103us -> 101us 0.99x
bitmap_RGBA_8888_A_scale_rotate_bilerp 112us -> 109us 0.98x
bitmap_RGBA_8888_update_volatile 5.72us -> 5.58us 0.98x
bitmap_RGBA_8888 5.73us -> 5.58us 0.97x
bitmap_RGBA_8888_update 5.78us -> 5.6us 0.97x
bitmap_RGBA_8888_A_scale_bilerp 70.7us -> 68us 0.96x
bitmap_RGBA_8888_A_scale_bicubic 23.7us -> 21.8us 0.92x
bitmap_RGBA_8888_A 13.9us -> 10.9us 0.78x
bitmap_RGBA_8888_A_source_opaque 14us -> 6.29us 0.45x
bitmap_RGBA_8888_A_source_transparent 14us -> 3.65us 0.26x
Running over our ~70 SKP web page captures, this looks like we spend 0.7x
the time in S32A_Opaque_BlitRow compared to the SSE2 version, which should
be a decent predictor of real-world impact.
BUG=chromium:399842
Review URL: https://codereview.chromium.org/874863002
Adds optimization of Skia S32A_Opaque_Blitrow blitter using SSE4.2 SIMD
instruction set. Special case for when alpha is zero or opaque.
Performance increase of 10%-400% compared to the existing SSE2
optimization (measured on Silvermont architecture).
Noticeable in ~25 different skia bench subtests, especially in
bitmap_8888_*, repeatTile_*, and morph_*.
bitmap_8888_A - 100% faster
bitmap_8888_A_source_transparent - 250% faster
bitmap_8888_A_source_opaque - 25% faster
bitmap_8888_A_scale_bicubic - 75% faster
Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
R=reed@google.com, mtklein@google.com, tomhudson@google.com, djsollen@google.com, joakim.landberg@intel.com
Author: henrik.smiding@intel.com
Review URL: https://codereview.chromium.org/289473009
Aarch64 support
This change contains the necessary modifications to have Skia build and
run properly on an ARMv8 processor in aarch64 execution state.
Here's a list of the changes:
- add an arm64 target to the build system + SK_CPU_ARM64 flag
- MatrixTest was failing when built in Release mode. Fused MAC
instructions were generated which made some intermediate results
more accurate. As the test relies on result comparison, the more
precise results when compared to others led to a gap bigger than
what was tolerated. As I don't know if some actual skia code relies
on results being comparable, I've disabled fused MAC instruction
with -ffp-contract=off for arm64.
- Modify include/core/SkOnce.h to have barriers work.
- SK_CPU_ARM64 implies SK_ARM_NEON_MODE_ALWAYS.
- use existing Xfermode optimisations with modifications that can be
removed in the future when toolchains are ready. Also save a few
instructions is two Xfermodes (will apply to ARM too).
- use existing SkBoxBlur and SkMorphology optimisations.
- use existing SkBlitMask optimisations
- use existing BitmapProcState and Convolution optimisations.
Future changes will include:
- Blitters (only partialy merged upstream)
- SkUtils (there's little value in sending asm optimisations without
having them benchmarked on real hardware).
Signed-off-by: Kevin PETIT <kevin.petit@arm.com>
BUG=skia:
Committed: http://code.google.com/p/skia/source/detail?r=13980R=djsollen@google.com, reed@google.com, mtklein@google.com, halcanary@google.com
Author: kevin.petit@arm.com
Review URL: https://codereview.chromium.org/143423004
git-svn-id: http://skia.googlecode.com/svn/trunk@14025 2bbb7eff-a529-9590-31e7-b0007b416f81
This patch implements basics for Xfermode SSE optimization. Based on
these basics, SSE2 implementation of multiply_modeproc is provided. SSE2
implementation for other modes will come in future. With this patch
performance of Xfermode_Multiply will improve about 45%. Here are the
data on desktop i7-3770.
before:
Xfermode_Multiply 8888: cmsecs = 33.30 565: cmsecs = 45.65
after:
Xfermode_Multiply 8888: cmsecs = 17.18 565: cmsecs = 24.87
BUG=
R=mtklein@google.com
Author: qiankun.miao@intel.com
Review URL: https://codereview.chromium.org/202903004
git-svn-id: http://skia.googlecode.com/svn/trunk@14006 2bbb7eff-a529-9590-31e7-b0007b416f81
Reason for revert:
GYP's failing on most (all?) bots.
Original issue's description:
> ARM Skia NEON patches - 35 - First AArch64 support
>
> Aarch64 support
>
> This change contains the necessary modifications to have Skia build and
> run properly on an ARMv8 processor in aarch64 execution state.
>
> Here's a list of the changes:
>
> - add an arm64 target to the build system + SK_CPU_ARM64 flag
>
> - MatrixTest was failing when built in Release mode. Fused MAC
> instructions were generated which made some intermediate results
> more accurate. As the test relies on result comparison, the more
> precise results when compared to others led to a gap bigger than
> what was tolerated. As I don't know if some actual skia code relies
> on results being comparable, I've disabled fused MAC instruction
> with -ffp-contract=off for arm64.
>
> - Modify include/core/SkOnce.h to have barriers work.
>
> - SK_CPU_ARM64 implies SK_ARM_NEON_MODE_ALWAYS.
>
> - use existing Xfermode optimisations with modifications that can be
> removed in the future when toolchains are ready. Also save a few
> instructions is two Xfermodes (will apply to ARM too).
>
> - use existing SkBoxBlur and SkMorphology optimisations.
>
> - use existing SkBlitMask optimisations
>
> - use existing BitmapProcState and Convolution optimisations.
>
> Future changes will include:
>
> - Blitters (only partialy merged upstream)
>
> - SkUtils (there's little value in sending asm optimisations without
> having them benchmarked on real hardware).
>
> Signed-off-by: Kevin PETIT <kevin.petit@arm.com>
>
> BUG=skia:
>
> Committed: http://code.google.com/p/skia/source/detail?r=13980R=djsollen@google.com, reed@google.com, halcanary@google.com, kevin.petit@arm.comTBR=djsollen@google.com, halcanary@google.com, kevin.petit@arm.com, reed@google.com
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Author: mtklein@google.com
Review URL: https://codereview.chromium.org/216113005
git-svn-id: http://skia.googlecode.com/svn/trunk@13983 2bbb7eff-a529-9590-31e7-b0007b416f81
Aarch64 support
This change contains the necessary modifications to have Skia build and
run properly on an ARMv8 processor in aarch64 execution state.
Here's a list of the changes:
- add an arm64 target to the build system + SK_CPU_ARM64 flag
- MatrixTest was failing when built in Release mode. Fused MAC
instructions were generated which made some intermediate results
more accurate. As the test relies on result comparison, the more
precise results when compared to others led to a gap bigger than
what was tolerated. As I don't know if some actual skia code relies
on results being comparable, I've disabled fused MAC instruction
with -ffp-contract=off for arm64.
- Modify include/core/SkOnce.h to have barriers work.
- SK_CPU_ARM64 implies SK_ARM_NEON_MODE_ALWAYS.
- use existing Xfermode optimisations with modifications that can be
removed in the future when toolchains are ready. Also save a few
instructions is two Xfermodes (will apply to ARM too).
- use existing SkBoxBlur and SkMorphology optimisations.
- use existing SkBlitMask optimisations
- use existing BitmapProcState and Convolution optimisations.
Future changes will include:
- Blitters (only partialy merged upstream)
- SkUtils (there's little value in sending asm optimisations without
having them benchmarked on real hardware).
Signed-off-by: Kevin PETIT <kevin.petit@arm.com>
BUG=skia:
R=djsollen@google.com, reed@google.com, mtklein@google.com, halcanary@google.com
Author: kevin.petit@arm.com
Review URL: https://codereview.chromium.org/143423004
git-svn-id: http://skia.googlecode.com/svn/trunk@13980 2bbb7eff-a529-9590-31e7-b0007b416f81
