| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
| |
We can record multiple frames in an .skp by recording SkCanvas::flush().
This should make SkPictures, SkLiteDL, and .skp files all record flush().
Change-Id: I6cf6e0e4ef993530d9f92fa168a53702ffce7d5e
Reviewed-on: https://skia-review.googlesource.com/34081
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This CL keeps the impl for each device backend in the
utils file for simplicity (shared helpers). Future CLs
may move into their respective impl as they become
more specialized.
Bug: skia:
Change-Id: I97ce6cdcc5106ebf4c84778f943cc32d0b7613c1
Reviewed-on: https://skia-review.googlesource.com/15893
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
| |
Bug: skia:6557
Change-Id: I0dbf70c4131ab59e7fc6c674a6587767af98e13a
Reviewed-on: https://skia-review.googlesource.com/15151
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
|
|
|
|
|
|
|
|
|
| |
Bug: skia:6557
Change-Id: I6482d74be7b360c93141a73dd80c67854530c7a1
Reviewed-on: https://skia-review.googlesource.com/15101
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
SkRecorder has some private Skia behavior with the way it handles an
SkNoDrawCanvas that is not exposed. In particular, it's able to create
an arbitrary SkIRect device clip at construction time.
This is a performance optimization when reusing SkNoDrawCanvas across
multiple recordings instead of reconstructing a new canvas every time.
https://codereview.chromium.org/2768143002 is trying to implement
another version of SkPictureBuilder/SkRecorder and so wants to be able
to use similar SkNoDrawCanvas behavior.
Change-Id: I0ac5f0eef8f0548cefd7354062311667cf448b07
Reviewed-on: https://skia-review.googlesource.com/11320
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
SkBitmapDevice-only implementation.
Will add A8 fast path specializations in a follow-up.
Change-Id: I2ccb1ffba3689e92ac90a23e94737471dfb121a1
BUG=skia:6005
Change-Id: I2ccb1ffba3689e92ac90a23e94737471dfb121a1
Reviewed-on: https://skia-review.googlesource.com/14183
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>
|
|
|
|
|
|
|
|
|
| |
BUG=skia:6366
Change-Id: Ic422fa44a788d3488c050c6218dbfba188bb8f3e
Reviewed-on: https://skia-review.googlesource.com/9835
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The fix was to release the array of vertices in the picturerecorder
destructor (where we also release textblobs etc.
This reverts commit 1eb3fef136bc75bd8e8ed717ec7c5d4ab26def62.
BUG=skia:
Change-Id: I3bf4acd6ad209205b0832a3cb7f94cd89dfcefc5
Reviewed-on: https://skia-review.googlesource.com/9826
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 14583e11fd622c686993b741499060a6f3527055.
Reason for revert: leaking
Direct leak of 499104 byte(s) in 2112 object(s) allocated from:
#0 0x1e195f0 in operator new(unsigned long) (/b/swarm_slave/w/irazbR79/out/Debug/dm+0x1e195f0)
#1 0x3142b0a in SkVertices::Builder::init(SkCanvas::VertexMode, int, int, SkVertices::Sizes const&) (/b/swarm_slave/w/irazbR79/out/Debug/dm+0x3142b0a)
Original change's description:
> More SkVertices implementation work
>
> - change virtuals to take const SkVertices*, as we do for TextBobs and Images
> - override onDrawVerticesObject in recording canvases
> - deserialize raw-vertices into SkVertices object
>
> Possibly a follow-on would intercept the raw-form directly in canvas,
> and remove the virtual, and only support the object form.
>
> BUG=skia:6366
>
> Change-Id: I57a932667ccb3b3b004beb802ac3ae6898e3c6e0
> Reviewed-on: https://skia-review.googlesource.com/9633
> Commit-Queue: Mike Reed <reed@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
>
TBR=bsalomon@google.com,reed@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:6366
Change-Id: I40bb7a20698ef6aa0a9ef71a3d6ac4c1473e081c
Reviewed-on: https://skia-review.googlesource.com/9825
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- change virtuals to take const SkVertices*, as we do for TextBobs and Images
- override onDrawVerticesObject in recording canvases
- deserialize raw-vertices into SkVertices object
Possibly a follow-on would intercept the raw-form directly in canvas,
and remove the virtual, and only support the object form.
BUG=skia:6366
Change-Id: I57a932667ccb3b3b004beb802ac3ae6898e3c6e0
Reviewed-on: https://skia-review.googlesource.com/9633
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
|
|
|
|
|
|
|
|
|
| |
Change-Id: Ieaa849bceba0e98e4c99491c721fe945a0694e68
Reviewed-on: https://skia-review.googlesource.com/9111
Reviewed-by: Mike Klein <mtklein@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Stan Iliev <stani@google.com>
Commit-Queue: Derek Sollenberger <djsollen@google.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit a129dfef2aaab0b5995cdf1ab7b2cdd41c29cf72.
BUG=skia:
Change-Id: I717de6e5fcd4516aa684b014b1414b0f82ac2b91
Reviewed-on: https://skia-review.googlesource.com/5722
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 8e7432b7f98dd592e529a0c8bb038d73ebfb0478.
Reason for revert: <INSERT REASONING HERE>
external/skia/bench/../tools/android/SkAndroidSDKCanvas.h:103:36: error: C++ requires a type specifier for all declarations
void onClipRect(const SkRect&, ClipOp, ClipEdgeStyle) override;
Original change's description:
> remove SK_SUPPORT_LEGACY_CLIP_REGIONOPS
>
>
> switch over to SkClipOps now that SK_SUPPORT_LEGACY_CLIP_REGIONOPS is gone
>
> BUG=skia:
>
> Change-Id: Ifdc8b3746d508348a40cc009a4e529a1cb3c405d
> Reviewed-on: https://skia-review.googlesource.com/5714
> Commit-Queue: Mike Reed <reed@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
>
TBR=reed@google.com,reviews@skia.org
BUG=skia:
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: If26ea91d7464615e43c1d3d2f726e337ff56b55c
Reviewed-on: https://skia-review.googlesource.com/5721
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
|
|
|
| |
switch over to SkClipOps now that SK_SUPPORT_LEGACY_CLIP_REGIONOPS is gone
BUG=skia:
Change-Id: Ifdc8b3746d508348a40cc009a4e529a1cb3c405d
Reviewed-on: https://skia-review.googlesource.com/5714
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
|
| |
R=reed@google.com,mtklein@google.com
Change-Id: I7df84587821db561e262a915afa23108a4a78834
Reviewed-on: https://skia-review.googlesource.com/5508
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
| |
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4384
Change-Id: I5121acd027d935ade169ff65941f29f654a47bd0
Reviewed-on: https://skia-review.googlesource.com/4384
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4020
CQ_INCLUDE_TRYBOTS=master.client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot
Change-Id: I19cd056f2af778f10e8c6c2b7b2735593b43dbac
Reviewed-on: https://skia-review.googlesource.com/4020
Reviewed-by: Florin Malita <fmalita@chromium.org>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Reed <reed@google.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This shouldn't be any significant change in behavior or performance,
but it leaves the code in SkLiteDL a bit simpler.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2863
Change-Id: Ifb0023d56f7e9dae172d8021fb956b8fc9281b15
Reviewed-on: https://skia-review.googlesource.com/2863
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
| |
SkRegion::Op --> SkCanvas::ClipOp (alias) --> SkClipOp
pre-CL needed in chrome : https://codereview.chromium.org/2355583002/
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2355483002
Review-Url: https://codereview.chromium.org/2355483002
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
of https://codereview.chromium.org/2227833004/ )
Reason for revert:
This CL breaks Android CTS tests CanvasStateTests#testClipPathReturnValues, CanvasStateTests#testClipRectReturnValues, and CanvasStateTests#testQuickReject.
Reverting it won't significantly affect Android's performance. See the following test result of testInvalidateTree Jank test. (Derek said that this is the only test closely related to this CL.)
**After reverting**
INSTRUMENTATION_STATUS: gfx-max-slow-draw=0.0
INSTRUMENTATION_STATUS: gfx-max-high-input-latency=0.0
INSTRUMENTATION_STATUS: gfx-max-slow-ui-thread=100.0
INSTRUMENTATION_STATUS: gfx-avg-slow-bitmap-uploads=100.0
INSTRUMENTATION_STATUS: gfx-avg-missed-vsync=72.96467105972435
INSTRUMENTATION_STATUS: gfx-avg-high-input-latency=0.0
INSTRUMENTATION_STATUS: gfx-max-slow-bitmap-uploads=100.0
INSTRUMENTATION_STATUS: gfx-max-missed-vsync=73.56181150550796
INSTRUMENTATION_STATUS: gfx-avg-frame-time-90=69.0
INSTRUMENTATION_STATUS: gfx-avg-frame-time-95=69.0
INSTRUMENTATION_STATUS: gfx-avg-frame-time-99=69.0
INSTRUMENTATION_STATUS: gfx-max-jank=100.0
INSTRUMENTATION_STATUS: gfx-avg-slow-draw=0.0
INSTRUMENTATION_STATUS: gfx-avg-slow-ui-thread=100.0
INSTRUMENTATION_STATUS: gfx-max-frame-time-90=69
INSTRUMENTATION_STATUS: gfx-max-frame-time-95=69
INSTRUMENTATION_STATUS: gfx-max-frame-time-99=69
INSTRUMENTATION_STATUS: gfx-avg-jank=100.0
INSTRUMENTATION_STATUS_CODE: -1
.
Test results for InstrumentationTestRunner=.
Time: 77.404
**Before Reverting**
gfx-max-slow-draw=1.1135857461024499
INSTRUMENTATION_STATUS: gfx-max-high-input-latency=0.0
INSTRUMENTATION_STATUS: gfx-max-slow-ui-thread=100.0
INSTRUMENTATION_STATUS: gfx-avg-slow-bitmap-uploads=100.0
INSTRUMENTATION_STATUS: gfx-avg-missed-vsync=70.79687591131983
INSTRUMENTATION_STATUS: gfx-avg-high-input-latency=0.0
INSTRUMENTATION_STATUS: gfx-max-slow-bitmap-uploads=100.0
INSTRUMENTATION_STATUS: gfx-max-missed-vsync=76.94174757281553
INSTRUMENTATION_STATUS: gfx-avg-frame-time-90=67.66666666666667
INSTRUMENTATION_STATUS: gfx-avg-frame-time-95=67.66666666666667
INSTRUMENTATION_STATUS: gfx-avg-frame-time-99=69.0
INSTRUMENTATION_STATUS: gfx-max-jank=100.0
INSTRUMENTATION_STATUS: gfx-avg-slow-draw=0.5317509395832639
INSTRUMENTATION_STATUS: gfx-avg-slow-ui-thread=100.0
INSTRUMENTATION_STATUS: gfx-max-frame-time-90=69
INSTRUMENTATION_STATUS: gfx-max-frame-time-95=69
INSTRUMENTATION_STATUS: gfx-max-frame-time-99=69
INSTRUMENTATION_STATUS: gfx-avg-jank=100.0
INSTRUMENTATION_STATUS_CODE: -1
.
Test results for InstrumentationTestRunner=.
Time: 77.411
OK (1 test)
Original issue's description:
> SkLiteRecorder: don't tell SkCanvas about clips
>
> If no one reads our clip, and we don't expect quickReject() to help,
> we can probably get away without maintaining a proper clip stack.
>
> This puts us at about 0.6-0.7x of previous record cost.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2227833004
>
> Committed: https://skia.googlesource.com/skia/+/8369e32a05543c2619b1d7e740fe6ff6d2f03af3
TBR=reed@google.com,mtklein@google.com,mtklein@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=skia:
Review-Url: https://codereview.chromium.org/2335773002
|
|
|
|
|
|
|
| |
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2303033002
Review-Url: https://codereview.chromium.org/2303033002
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This will allow us to optimize for the RectGrid macrobench.
Currently, SkiaGL is much slower than OpenGL.
SkiaGL 12 items/s
OpenGL 160 items/s
This contains everything except for the fast implementation on GPU.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2277053002
Review-Url: https://codereview.chromium.org/2277053002
|
|
|
|
|
|
|
| |
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2224163005
Review-Url: https://codereview.chromium.org/2224163005
|
|
|
|
|
|
|
|
|
| |
Plumbs the drawArc canvas method down to SkDevice without converting to a path. Plumbs through the various recording canvas classes.
BUG=skia:5227
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2257023003
Review-Url: https://codereview.chromium.org/2257023003
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds didTranslate() so that SkLiteDL (and other canvas recorders)
can record the translate rather than the full concat.
It also adds a case to SkMatrix::preTranslate() to fast path
translate x translate -> translate (i.e. +=).
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255283002
Committed: https://skia.googlesource.com/skia/+/5fa47f4fd13b3158de4599414c86d17649c2dd1c
Misc bots failing in pictureimagefilter replay modes.
https://luci-milo.appspot.com/swarming/task/30b8e53f3a1f4f10/steps/dm/0/stdout
Problem is FMA vs. not.
CQ_INCLUDE_TRYBOTS=master.client.skia:
Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-Fast-Trybot
Review-Url: https://codereview.chromium.org/2255283002
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
id:80001 of https://codereview.chromium.org/2255283002/ )
Reason for revert:
speculative
Original issue's description:
> Fast path translate() in SkCanvas and SkLiteDL.
>
> This adds didTranslate() so that SkLiteDL (and other canvas recorders)
> can record the translate rather than the full concat.
>
> It also adds a case to SkMatrix::preTranslate() to fast path
> translate x translate -> translate (i.e. +=).
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255283002
>
> Committed: https://skia.googlesource.com/skia/+/5fa47f4fd13b3158de4599414c86d17649c2dd1c
TBR=herb@google.com,reed@google.com,mtklein@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review-Url: https://codereview.chromium.org/2264433002
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds didTranslate() so that SkLiteDL (and other canvas recorders)
can record the translate rather than the full concat.
It also adds a case to SkMatrix::preTranslate() to fast path
translate x translate -> translate (i.e. +=).
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2255283002
Review-Url: https://codereview.chromium.org/2255283002
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
out/Release/nanobench --match Lattice --config gpu --ms 3000
3.42ms -> 17.2us
For reference, a loop over drawBitmapRects (which is what
Android currently does) is about 13us.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2205273003
Review-Url: https://codereview.chromium.org/2205273003
|
|
|
|
|
|
|
|
|
|
|
|
| |
If no one reads our clip, and we don't expect quickReject() to help,
we can probably get away without maintaining a proper clip stack.
This puts us at about 0.6-0.7x of previous record cost.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2227833004
Review-Url: https://codereview.chromium.org/2227833004
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fix saveLayer bounds
all the misc calls
a bit of refactoring
just drawVertices and drawPatch to go, plus some todos and likely bug fixes
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2226563002
Review-Url: https://codereview.chromium.org/2226563002
|
|
|
|
|
|
|
|
|
|
|
| |
- optimizeFor(GrContext*)
- images
- basic text
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2224753002
Review-Url: https://codereview.chromium.org/2224753002
|
|
SkLiteRecorder, a new SkCanvas, fills out SkLiteDL, a new SkDrawable.
This SkDrawable is a display list similar to SkRecord and SkBigPicture / SkRecordedDrawable, but with a few new design points inspired by Android and slimming paint:
1) SkLiteDL is structured as one big contiguous array rather than the two layer structure of SkRecord. This trades away flexibility and large-op-count performance for better data locality for small to medium size pictures.
2) We keep a global freelist of SkLiteDLs, both reusing the SkLiteDL struct itself and its contiguous byte array. This keeps the expected number of mallocs per display list allocation <1 (really, ~0) for cyclical use cases.
These two together mean recording is faster. Measuring against the code we use at head, SkLiteRecorder trends about ~3x faster across various size pictures, matching speed at 0 draws and beating the special-case 1-draw pictures we have today. (I.e. we won't need those special case implementations anymore, because they're slower than this new generic code.) This new strategy records 10 drawRects() in about the same time the old strategy took for 2.
This strategy stays the winner until at least 500 drawRect()s on my laptop, where I stopped checking.
A simpler alternative to freelisting is also possible (but not implemented here), where we allow the client to manually reset() an SkLiteDL for reuse when its refcnt is 1. That's essentially what we're doing with the freelist, except tracking what's available for reuse globally instead of making the client do it.
This code is not fully capable yet, but most of the key design points are there. The internal structure of SkLiteDL is the area I expect to be most volatile (anything involving Op), but its interface and the whole of SkLiteRecorder ought to be just about done.
You can run nanobench --match picture_overhead as a demo. Everything it exercises is fully fleshed out, so what it tests is an apples-to-apples comparison as far as recording costs go. I have not yet compared playback performance.
It should be simple to wrap this into an SkPicture subclass if we want.
I won't start proposing we replace anything old with anything new quite yet until I have more ducks in a row, but this does look pretty promising (similar to the SkRecord over old SkPicture change a couple years ago) and I'd like to land, experiment, iterate, especially with an eye toward Android.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2213333002
Review-Url: https://codereview.chromium.org/2213333002
|