N=15 and N=63 make for nice even looking profiles
on ARM and x86 respectively, with N=15 running 3
body loops and 3 tail loops on ARM, N=63 running
7 body loops and 7 tail loops on x86.
Change-Id: Ie7616bd99c949328bbb7d7048fc6f468ff1e3ad2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/227220
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
These didn't work correctly, and they're extremely tricky to get right
in the vectorized execution model (vs. structured control flow). As a
side effect, determine the maximum stack depth used for the execution
masking - the same idea will be used for the primary stack in a later
CL. Add a unit test to verify the new restriction, and fix two places
that were relying on this feature before.
In addition, boolean external values need to be masks. I may implement
this in the code-gen at some point, but this is already a fringe
feature, so just fix the one unit test for now.
Change-Id: I9607ffaf67c7795dbf42e4009180aea8f3e65c44
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/226849
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
+ fix bug in SkBulkGlyphMetricsAndImages where the underlying SkAutoSTArray
was not reset.
Change-Id: I94882da30e858d012aa02f79c5fdfe4bc7d2e8aa
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/225541
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
config is completely redundant. No caller really cares what the backend
format is.
Change-Id: I93f1feb3ee61db6c21b7915bab3ee3fba5656f92
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/225194
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
I'm kind of getting bored of having to pass --loops 0
in optimized+asserts builds.
We'd been defaulting to 1 loop in Debug and ASAN builds,
and manually setting Valgrind builds to 1 loop with 1
sample. Remove the default so all builds auto-tune,
but set Debug and ASAN bots like Valgrind bots.
Change-Id: Ifedd98a11ea74a2c222fe59bd50b96683d12b98c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/224816
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
There are a few places that have been checking whether the *proxy*
needs stencil, in order to determine if the current render target
context needs stencil. This is problematic since a render target
context can not require stencil itself, but wrap an existing proxy
that already has stencil.
Bug: skia:
Change-Id: I2719dd3a9df15fef3d64f991cda4fadea23266bb
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/223970
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Because it mutates the glyph.
Change-Id: Ic7ce320350764454d7a76335828d398f19b149d6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/223797
Commit-Queue: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This is largely redundant with GrPixelConfig. However, we intend to
remove GrPixelConfig.
Bug: skia:7580
Change-Id: I03d92303be832711f7821f8a97d36387c9b04a9f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/222883
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Looks like ~50ns overhead for RP vs ~14,000ns for SkVM.
Change-Id: I85ef73d3387657b14615fcfa5cfd9df5c2325343
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/223302
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Converting to glyph() style calls that return SkGlyph*. This is mainly preparation
for removing converting findImage(const SkGlyph&) to prepareImage(SkGlyph*).
+ Misc cleanups mainly fWidth -> width() type things.
Change-Id: Id5c9b0ba5856b3ea54353ece4d05fa495cc5a640
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/223187
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
From now on, sample counts always refer to the number of actual color
samples, and render targets don't have separate color and stencil
sample counts.
If mixed samples support is available when making a
"GrAAType::kCoverage" draw, then an op may attach and use a mixed
sampled stencil buffer internally. But this will all be invisible to
the client.
After this CL, we temporarily won't have a mode to use nvpr with mixed
samples. That will soon be fixed by a follow-on CL that enables nvpr
with mixed samples in the normal "gl" and "gles" configs.
Bug: skia:
Change-Id: I1cb8277f0d2d0d371f24bb9f39cd473ed5c5c83b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221878
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This new bench lets us measure the overhead of program building,
optimization, and JITting. Surprisingly, at head the optimization in
Builder::done() takes longer than the JIT.
The new bench clocks in around 40µs on my laptop at head,
then 32µs after switching val_to_reg to be an std::vector,
then 27µs after switching deaths to be an std::vector too,
then 22µs after switching fIndex to be an SkTHashMap,
then 20µs after calling program.reserve(fProgram.size()),
then 19µs after switching JIT data maps to SkTHashMap too.
I tried swapping some std::vector for SkTDArray to no benefit, actually
a little detriment. So I think this is roughly all the low-hanging
fruit, with time split now roughly equally between Builder::Done(),
JITting in Program::eval(), and the original calls to Builder
themselves.
Also disable perf dumps on Mac. No real value there until I can dump a
dylib, and it's just one more thing I have to remember to disable before
running this sort of benchmark.
Change-Id: I1c6e58ed00ac94ad622c7d740712634f60787102
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/222984
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This moves the responsibility for allocating executable code out of
Assembler. The pages Xbyak uses are obviously executable, so this is
redundant right now, but it'll let us switch to something simple like
std::vector<uint8_t> as we continue to cut out Xbyak.
Make how Program holds its cached JIT program slightly less of a mess.
Change-Id: I38d6f01006da1da60f4aed675e9ddf97de9aec52
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/222575
Auto-Submit: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
- 32x8 i32 add,sub,mul
- add I32_Naive bench/test builder to get better i32 mul coverage
- minor refactoring all over
Change-Id: I13cc19ff37a2da0bcff289ba51baac08f456d6c5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/222485
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This extends the runStriped to all for return data.
GPU impl not done yet, will be done in a follow-on CL.
Change-Id: Ib107d2945f6fdb34ce1b5405a6c88a5ae7e9f7ac
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221539
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Reed <reed@google.com>
This reverts commit 59b22f9297.
Reason for revert: ios
Original change's description:
> Reland "Reland "Interpreter: Support striped inputs for less overhead""
>
> This reverts commit 2c59b4e9ea.
>
> Change-Id: I2b06936994430722b8fc3890ff9b4a6f4710db04
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221998
> Reviewed-by: Brian Osman <brianosman@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
TBR=mtklein@google.com,brianosman@google.com,reed@google.com
Change-Id: I321d257e400f5df3c6115dfe9aa18a2794a95910
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/222099
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit 2c59b4e9ea.
Change-Id: I2b06936994430722b8fc3890ff9b4a6f4710db04
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221998
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit edc42b9971.
Reason for revert: Sigh
Original change's description:
> Revert "Revert "Interpreter: Support striped inputs for less overhead""
>
> This reverts commit 645fe10313.
>
> Change-Id: If74a15479f89f49ac33c0b6241bb0db92bc11083
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221721
> Reviewed-by: Mike Klein <mtklein@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
TBR=mtklein@google.com,brianosman@google.com,reed@google.com
Change-Id: I651d56821bbde8b91887aa885bcf2cb202707388
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221897
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit 5d89b66ff6.
Reason for revert: Chrome roll
Original change's description:
> Interpreter: Support striped inputs for less overhead
>
> Change-Id: I8c7bd5ed3fb6aebbfb1c5c224acfd73862252621
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/220778
> Commit-Queue: Brian Osman <brianosman@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
TBR=mtklein@google.com,brianosman@google.com,ethannicholas@google.com,reed@google.com
Change-Id: I2275e7ed12031e58e59b6618d0ea59bec2ee2815
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221716
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Change-Id: I8c7bd5ed3fb6aebbfb1c5c224acfd73862252621
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/220778
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Reed <reed@google.com>
In GrRecordingContext I moved the auditTrail onto the heap and only there
when compiling for tests. This allowed us to move a lot of files out of
include private.
Change-Id: Ib76ac211c0c6fd10bacaccf0c5f93f21a59f35d5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/221344
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Auto-Submit: Greg Daniel <egdaniel@google.com>
Slightly sharper, but far easier to hold:
- Remove Value union from interface, everything is a 32-bit
value type, or a collection thereof.
- Collapse to one version of Run (that takes count), and make
it a member on ByteCode.
- Similarly, move disassemble to ByteCodeFunction.
Change-Id: I07c85e65991178b3f52e20e815c25f36bc9c4257
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/220753
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This converts the SkSL interpreter to operate in SIMT fashion. It handles
all the same features as the previous scalar implementation, but operates
on N lanes at a time. (Currently 8).
It's modeled after GPU and other parallel architectures, using execution
masks to handle control flow, including divergent control-flow.
Change-Id: Ieb38ffe2f55a10f72bdab844c297126fe9bedb6c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217122
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This is a reland of 10ad0b9b01
Original change's description:
> SkParagraph
>
> Change-Id: I0a4be75fd0c18021c201bcc1edfdfad8556edeff
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/192100
> Reviewed-by: Ben Wagner <bungeman@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
> Commit-Queue: Julia Lavrova <jlavrova@google.com>
Change-Id: I46cf43eae693edf68e45345acd0eb39e04e02bfc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/219863
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Julia Lavrova <jlavrova@google.com>
Add MakeWithNoDevice to avoid the threshold of converting to
paths when using MakeCanonicalized.
Change-Id: I73b49dd35ffb2f4fa3b70ba3324bfcb348ca913b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/218960
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
SkStrikeSpecStorage was a temporary name until all uses of
SkStrikeSpec were cleaned up. Do the final rename.
Change-Id: Iaba987ecdfe46ca9eee8d530d5095840cdca300d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/219209
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Bug: skia:
Change-Id: I83e3a094d26085fc4d586e5d2581e0d61c55634e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/145080
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Eliminate the duplicate functionality,
and better testing for the bench builders.
Change-Id: If20e52107738903f854aec431416e573d7a7d640
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/218041
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Convert over remaining callsights for strike lookup, and reduce the SkStrikeCache
API.
- one friend decl.
Change-Id: I109a747061030f26ff908472ea55a4424d5c10f1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/209109
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Change-Id: I864d3c2452f3affdc744bf8b11ed3b3e37d6d922
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/216602
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Things were running suspiciously well...
_I32 had a typo that cut out 3/4 of its multiplies...
_I32_SWAR was missing a mask operation needed to drop
the junk low byte of the high half after the multiply.
The bench times now make a bit more sense and are in line
with how much work we're actually doing: F32's the slowest,
I32 a little faster, and I32_SWAR fastest:
curr/maxrss loops min median mean max stddev samples config bench
35/36 MB 58 2.03ns 2.04ns 2.04ns 2.04ns 0% ▂▂▂▂▁▁█▁▂▁ nonrendering SkVM_4096_I32_SWAR
35/36 MB 42 3.44ns 3.48ns 3.49ns 3.59ns 1% ▂▆▅█▃▃▁▂▂▄ nonrendering SkVM_4096_I32
35/36 MB 30 4.9ns 5.21ns 5.11ns 5.33ns 3% ▆▇█▆▆▁▂▁▁▅ nonrendering SkVM_4096_F32
35/36 MB 203 0.696ns 0.697ns 0.705ns 0.758ns 3% █▂▂▁▁▁▁▁▁▂ nonrendering SkVM_4096_RP
35/36 MB 942 0.188ns 0.188ns 0.188ns 0.189ns 0% ▂▁▂▁▃█▂▁▁▁ nonrendering SkVM_4096_Opts
Change-Id: I2850dc3f9df1828f03499eb278b8231f48eaae63
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217982
Commit-Queue: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
With all the thinking around a stack-based interpreter,
I figured I'd sketch out some ideas for a register VM too.
I kind of have the hunch that this is the direction that
will actually let us replace large amounts of Skia's CPU
backend with an efficient interpreter or JIT.
Change-Id: Ia2b5ba4a3fc27556f5b6ba95cd1ace46d3217403
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/216665
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Bug: skia:
Change-Id: I8497ddbc5c5fef4e0fa87f3b9828034ba34284cc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217636
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
This lets you write a suite of benchmarks that
run against different sized workloads and see
numbers that normalized by that workload size,
e.g. nanoseconds / pixel.
C.f. the bench in "sketch an skvm".
Change-Id: I6106c51ceac59cd533449712a77c957bdf86846e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217416
Commit-Queue: Mike Klein <mtklein@google.com>
Auto-Submit: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This is a reland of a36e089065
This is only active when Metal is enabled.
Original change's description:
> Added AutoreleasePool for managing pool memory in testing apps.
>
> This is only active on MacOS and iOS -- on other platforms it
> will do nothing as they have no need for autorelease pools.
>
> Bug: skia:8243
> Change-Id: Ib74968dab6e3455a72e726429832101d0d410076
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217126
> Commit-Queue: Jim Van Verth <jvanverth@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
Bug: skia:8243
Change-Id: I743a3dcc93b46387a6a330e855c2e8810b482544
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217379
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
This reverts commit a36e089065.
Reason for revert: Primary suspect in breaking G3
Original change's description:
> Added AutoreleasePool for managing pool memory in testing apps.
>
> This is only active on MacOS and iOS -- on other platforms it
> will do nothing as they have no need for autorelease pools.
>
> Bug: skia:8243
> Change-Id: Ib74968dab6e3455a72e726429832101d0d410076
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217126
> Commit-Queue: Jim Van Verth <jvanverth@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
TBR=egdaniel@google.com,jvanverth@google.com,brianosman@google.com
Change-Id: I64f6e0baba21a9d35682ab53bdf418180be8579b
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:8243
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217377
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Kevin Lubick <kjlubick@google.com>
This is only active on MacOS and iOS -- on other platforms it
will do nothing as they have no need for autorelease pools.
Bug: skia:8243
Change-Id: Ib74968dab6e3455a72e726429832101d0d410076
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/217126
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Also some minor refactoring in make_image_shader().
Change-Id: I6061ebeb8c297045b1a276160fe18cc6ae54bde7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/216867
Commit-Queue: Hal Canary <halcanary@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Auto-Submit: Hal Canary <halcanary@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
This file pulls in Windows headers in a custom way, which is somewhat
awkward for a library header. The only use in include/ has been replaced
with a single forward declaration.
Change-Id: Ibef4cf7a2d1c9957a6a5b145b95aca1a6868cb5e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/214689
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Previous structure didn't measure anything but new/free cost. All of the
constructs turned into (eventually) drawRect + color.
New version also measures that, but in a clearer way, so we can examine
all the variants, and assume that they should all be about the same.
Change-Id: I1b4a0120a3e663ce1b9daa3c1e3e26ae6b397677
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/214687
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
It's been dead code since it was inlined into Android last summer.
Change-Id: I252f6392d9436ef357f22a54bab8a33c9d1b3ea9
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/213625
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>