This piece of code is already 64-bit only, so we don't need to think about ARMv7.
Hopefully this shuts up the warnings. They were harmless.
If this doesn't work (it's relatively new modifier, so maybe some compilers barf), an alternative is to cast count to a size_t.
BUG=skia:4686
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1527123003
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Review URL: https://codereview.chromium.org/1527123003
Mike points that the the ulps compares
rank high in path ops profiles. The
check for finite scalars is rarely
required.
Call it less by:
- specializing _pin version of compares
- checking for 0 divides up front
- handling failing cases before comparing
- casting float to double before adding
R=reed@google.com
Review URL: https://codereview.chromium.org/1522183002
This is mainly warmup for an AVX2 version.
The machine I'm typing this on just doesn't support AVX2.
This strategy should translate easily down to SSSE3 and SSE2.
Xfermode_SrcOver: 2.73ms -> 2.62ms (0.96x) (That's Color32.)
Xfermode_SrcOver_aa: 3.48ms -> 3.09ms (0.89x) (That's BlitMask_D32_A8.)
AA text blits (text_16_AA_{88,FF,WT,BK}) show speedups in the range of 5 to 20%.
Unlike previous versions of this code, all the div255() are exactly (x+127)/255.
This won't fix any major bugs, but it does correct our bias in the middle.
There will be many diffs, all minor.
I've punted for now on pmaddubsw for lerping. I do intend to try that,
but I want this (relatively simple) code as my basis for comparison.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1526883004
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Review URL: https://codereview.chromium.org/1526883004
These were originally used to compare to the old implementation of
subset decoding in SkImageDecoder. The old implementation has been
removed, and they do not provide useful information that is not
covered by the BitmapRegionDecoderBenches.
This will greatly speed up some of our infra bots, which spend a lot of
time decoding interlaced PNGs repeatedly (thanks to
SubsetTranslateBench).
BUG=skia:4715
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1531833002
Review URL: https://codereview.chromium.org/1531833002
We are trying to support the behavior provided by the Android
framework. Other developers do not necessarily need this behavior.
BUG=skia:4296
Review URL: https://codereview.chromium.org/1518933002
When the default framebuffer is wrapped in a device for rendering we don't get a GrTexture. This CL adds a copy to a temporary texture in this instance so the rest of the Ganesh pipeline can continue on as usual.
Review URL: https://codereview.chromium.org/1531493002
Fix case where texture format and data format are different.
Use helper when createing a testing "backend" texture.
BUG=skia:
Review URL: https://codereview.chromium.org/1527753002
This is more general than checking for __linux, __FreeBSD__, etc. (In
principle we could remove some of the existing checks such as
__FreeBSD__, but I have not tried that so far.)
In particular, it allows Skia to build with the NaCl or PNaCl
toolchains, which is something we would like for Mojo.
BUG=https://github.com/domokit/mojo/issues/431
TEST=none
Review URL: https://codereview.chromium.org/1523733003
Given the autovectorization we've seen, I wouldn't expect big speedups
from this, but it does give us a point of control over what's going on.
BUG=skia:
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Review URL: https://codereview.chromium.org/1526923003
Epsilon bias to keep bitmap sample rounding consistent with geometry
rounding.
Also update the GM to draw an outer border + drop uninteresting
scales in favor of negative scale variants.
BUG=skia:4680,skia:4649
R=reed@google.com,caryclark@google.com,mtklein@google.com
Review URL: https://codereview.chromium.org/1527633002
Reason for revert:
linux Canary builder has no std::steady_clock. Weird...
Original issue's description:
> SkTime updates
>
> 1) Use steady_clock instead of high_resolution_clock. If we don't have a
> guarantee of monotonicity, it's pretty much useless for timing things.
>
> 2) Implement Mac/iOS with <chrono> too. This was waiting on C++11 library support.
>
> Both high_resolution_clock and steady_clock are (still) busted on MSVC 2013,
> so no change there.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/6a20871e5aeaa7e61f3348694bf436af16f824b9TBR=herb@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/1529603002
Running `Release/dm --gpu 0`, the number of times we call
SkBitmap::operator=(const SkBitmap&)
(which refs the pixelref) is reduced from ~214929 to ~214626.
Review URL: https://codereview.chromium.org/1514503004
1) Use steady_clock instead of high_resolution_clock. If we don't have a
guarantee of monotonicity, it's pretty much useless for timing things.
2) Implement Mac/iOS with <chrono> too. This was waiting on C++11 library support.
Both high_resolution_clock and steady_clock are (still) busted on MSVC 2013,
so no change there.
BUG=skia:
Review URL: https://codereview.chromium.org/1521293002
