This reverts commit 6b2121d0ec.
Reason for revert: Perf regressions:
https://perf.skia.org/t/?begin=1623217407&end=1623217409&subset=all
Original change's description:
> Wean tessellation off SkPathOps
>
> Now that chopping and culling are fully implemented, and the indirect
> tessellators are gone, we don't need to crop huge paths anymore. The
> recursive chopping and culling will just do their thing.
>
> Bug: skia:9795
> Bug: skia:11268
> Bug: chromium:800804
> Change-Id: Ie34920c59af67035cf37eead69c5335693703001
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/416080
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
TBR=csmartdalton@google.com,reed@google.com
Change-Id: I265bd6991627b435a2a656cf9080b7e9d17e5ded
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:9795
Bug: skia:11268
Bug: chromium:800804
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/417238
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
We can't use our hardware tessellation back door if any FPs in the
program have varyings. Before this CL, we were forgetting to check the
clip FP for strokes, and weren't checking any FPs yet for fills.
Bug: skia:10419
Change-Id: Ica631ab3cf0407fb359c02c6d53f88f5f301cddc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/417237
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Now that chopping and culling are fully implemented, and the indirect
tessellators are gone, we don't need to crop huge paths anymore. The
recursive chopping and culling will just do their thing.
Bug: skia:9795
Bug: skia:11268
Bug: chromium:800804
Change-Id: Ie34920c59af67035cf37eead69c5335693703001
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/416080
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Bug: skia:10419
Change-Id: Icb3395565060d736624d03ba1465926bd9188e44
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/416078
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Rearranges the algebra to use "precision". This makes it consistent
with the other formulas and removes the need for the caller to think
about inverting its precision value.
Bug: skia:10419
Change-Id: I186d03393952983e86b0bef69e1f89f86cdbb423
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/414616
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Tyler Denniston <tdenniston@google.com>
It was becoming a bit chaotic to have so many different algorithms in
one file.
Bug: skia:10419
Change-Id: I477b7b65439d16b0f983e10952cb41ec881e8183
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/413856
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Merges the GrStrokePathShader and GrFillPathShader classes together.
Creates a new base class called GrPathTessellationShader. Now they all
have a uniform color and can all be draw to stencil and color both.
This is necessary cleanup in general, but will also allow us to create
a convex tessellation op that bypasses the stencil buffer.
Bug: skia:10419
Change-Id: Ifc492c94d3de044a36bd9ea95b1d5aa22e007905
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/413696
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
The fixed count stroke shader was executing a sort loop long enough to
handle 1024 parametric segments, when in reality it will never see
more than 48. This CL cuts the per-vertex sorting work nearly in half
by reducing the number of iterations from 10 to 6. (The indirect
tessellator continues to use 10 iterations.)
Bug: skia:10496
Change-Id: Idc21d6015d449f63915780972a8d487c9c6de6fe
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/412496
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Now that hardware tessellators chop, support raw triangles, and can
handle any path, we don't need complicated logic anymore to determine
when we can't use them. This CL simplifies the criteria for selecting
a tessellation algorithm and adds a fAlwaysPreferHardwareTessellation
context option to override it.
Bug: skia:10419
Change-Id: I8492e8f285ff27eb9d0dd6b1e9817dbeeb386c63
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/411496
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Previously we would completely disable hardware tessellation for a path
if there was any chance of a curve requiring more segments than
supported by the hardware. This CL updates the tessellators to simply
chop paths until they fit in patches, allowing us to finally draw any
path using hardware tessellation.
Bug: skia:10419
Change-Id: I5c9f78cda3e30b8810aff3cb908235965706f2d8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/410977
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Supports drawing exact triangles as conics with infinite weight. This
is necessary for drawing inner fans, breadcrumb triangles, and extra
triangles that will be introduced when we start splitting curves that
don't fit in a single patch.
Bug: skia:10419
Change-Id: If72560be4eb38340512be55dccea93b3a083fad5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/410836
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Splits the instanced and tessellation implementations into their own
separate files and merges their common code into a shared base class.
Renames GrStrokeTessellateShader to GrStrokeShader.
Bug: skia:10419
Change-Id: Ia731509858e682a605ee65c9ced1fd163e4c03f3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/409036
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Tyler Denniston <tdenniston@google.com>
This is a reland of 99e6f0fcfb
Original change's description:
> Reland "Reland "Use conics with w=Inf to describe triangles for the tessellator""
>
> This is a reland of ff515df5b4
>
> Original change's description:
> > Reland "Use conics with w=Inf to describe triangles for the tessellator"
> >
> > This is a reland of 84f70136ab
> >
> > Original change's description:
> > > Use conics with w=Inf to describe triangles for the tessellator
> > >
> > > Previously, only the indirect tessellator could draw triangles, and
> > > only with special index data. Using conics with w=Inf will allow us to
> > > draw triangles with the hardware tessellator as well, in addition to
> > > being able to wean the indirect tessellator off an index buffer.
> > >
> > > Bug: skia:10419
> > > Bug: chromium:1202607
> > > Change-Id: I180af9cb5410c0e0bb25a2edcfb01e17d4a2f590
> > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/406977
> > > Commit-Queue: Chris Dalton <csmartdalton@google.com>
> > > Reviewed-by: Michael Ludwig <michaelludwig@google.com>
> >
> > Bug: skia:10419
> > Bug: chromium:1202607
> > Change-Id: Ic12b10eaa60fddd212c66757bf7100749ee58d49
> > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408096
> > Reviewed-by: Robert Phillips <robertphillips@google.com>
> > Commit-Queue: Chris Dalton <csmartdalton@google.com>
>
> TBR=robertphillips@google.com
>
> Bug: skia:10419
> Bug: chromium:1202607
> Change-Id: Id1f8dfa133f446b9f30f2bf8493a3a7b99072e59
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408616
> Reviewed-by: Chris Dalton <csmartdalton@google.com>
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
Bug: skia:10419
Bug: chromium:1202607
Change-Id: Iadc60aae395146d5da5bfaa725b980c963e6bead
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408817
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
When a stroke is zoomed in, it can require an exponential number of
chops. This CL updates the tessellators to stop recursing once a curve
is no longer in the viewport.
Bug: skia:10419
Bug: skia:11268
Bug: skia:9795
Change-Id: I6aad5355618462e88ce6ccccb57c500ae6310f22
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408576
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Bug: skia:10419
Change-Id: I061c216e6f49399e533fb9a83ba5d9f9d3af83e3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/408858
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
- most of the small diffs are because I moved GrWangsFormula.h out
of the tessellate/ directory and into the geometry/ directory since
it's more general than HW tessellation.
The previous implementation was based on the heuristic that the distance
from the true curve to the line segment would be divided by 4 every time
the curve was recursively subdivided. This was a reasonable
approximation if the curve had balanced curvature on both sides of the
split. However, in the case of the new GM's curve, the left half was
already very linear and the right half had much higher curves.
This lead to the approximation reporting fewer points than required.
Theoretically, those few points that weren't utilized by the left half
of the curve could have been made available to the right half, but the
implementation of that would be tricky.
Instead, it now uses Wang's formula to compute the number of points.
Since recursive subdivision leads to linearly spaced samples assuming it
can't stop early, this point count represents a valid upper bound on
what's needed. It also then ensures both left and right halves of a
curve have the point counts they might need w/o updating the
generation implementations. However, since the recursive point
generation exits once each section has reached the error tolerance, in
scenarios where the prior approximation was reasonable, we'll end up
using fewer points than reported by Wang's. Hopefully that means there
is negligible performance regression since we won't be increasing
vertex counts by that much (except where needed for correctness).
Bug: skia:11886
Change-Id: Iba39dbe4de82011775524583efd461b10c9259fe
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/405197
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
This new tessellator renders strokes as fixed-count triangle strip
instances. Any extra triangles not needed by the instance are emitted
as degenerate triangles. Since it draws in order, this tessellator
allows us to batch dynamic colors even when hw tessellation is not
supported.
Bug: skia:10419
Change-Id: If03a8b76319471ae4d4580dda019b69204d9197b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398416
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Bug: skia:10419
Change-Id: I656e2508e95a47e03af431734ac06cbdb41232c6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/398356
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This will be used by the new stroke tessellator. All the other
tessellators should start chopping and chunking too. That will allow us
to quit cropping paths if we are afraid they might need more segments
than are supported.
Bug: chromium:1172543
Change-Id: I30f0ebb581f56cac099d8c05e0e181c4657c3db8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/390096
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Adds quick accepts to the switch statement in
GrStrokeHardwareTessellator::prepare() that allow us to write out most
tessellation patches immediately. This avoids making function calls
into recursive methods as well as avoiding some of their checks that
aren't necessary the first time around.
Also adds a microbench that mimics the MotionMark "paths" benchmark
and measures our CPU-side prepare() time.
This shaves up to 30% off the microbenchmarks.
Bug: chromium:1172543
Change-Id: Idc93bebb79db9898a4ec241b1f6c8b9eb9ba7da3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/372602
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
The intent of this class was to abstract the internal linked list used
by GrStrokeTessellateOp, but it seems to just make things more
complicated. We have a need now to iterate the list with more freedom
than is offered by GrSTArenaList, so it seems best to just use a plain
C-style linked list instead.
Bug: chromium:1172543
Change-Id: Ia76be83c523bd3c285200099a529ccd3818490b6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/372656
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
The indirect tessellator can't combine overlapping, mismatched colors
because the log2 binning draws things out of order. But we can still
chain them together and generate a single long list of indirect draws.
Bug: chromium:1172543
Bug: skia:10419
Change-Id: Id7fc7488411a2a189e24cd7dd692e5c78497f498
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/370197
Reviewed-by: Brian Salomon <bsalomon@google.com>
Only adds color to the hardware tessellator. The indirect tessellator
reorders draws with is log2 binning, so we can't have different
colors.
Bug: chromium:1172543
Bug: skia:10419
Change-Id: I2a3700cd4572e8222002bfb028af05c6ec447708
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/369976
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Allows us to batch together strokes that have different SkStrokeRecs.
Bug: chromium:1172543
Bug: skia:10419
Change-Id: I11dc01e60bc17a6bb3c3b635f9edb2944a2f2edc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/369579
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Also add a unit test that the vectorized version equals the reference
implementation.
Bug: skia:10419
Change-Id: I4d165fd45532e9ec468565d0637fb769b51f5fcd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345122
Commit-Queue: Tyler Denniston <tdenniston@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Now there is only one op to tessellate a stroke, and it creates its
own GrStrokeIndirectTessellator or GrStrokeHardwareTessellator
internally. This will allow us to dynamically switch into hardware
tessellation when we need to batch strokes that have different
parameters or colors.
Bug: chromium:1172543
Bug: skia:10419
Change-Id: I3cddb855fdbb9ab018785584497c843e3e31b75e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/366056
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Simplifies the op to always triangulate the inner fan. We ensure this
will always work by using breadcrumb triangles. Also removes the
fail-on-non-simple mode from GrInnerFanTriangulator since it isn't
being used anymore.
Bug: skia:10419
Change-Id: Idb849712bf2bc4e5ef785bc3f9b8db03119d230e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/359565
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
This reduces complexity and will also allow us to add new ops that
take advantage of this same core logic.
Bug: skia:10419
Change-Id: I4ec8717a6d9510dea967d11467eeea0b5b7c7f4c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/354296
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This is a reland of ad8b9fa2d7
Original change's description:
> Disable HW tessellation shaders by default
>
> The GLSL backdoor for tessellation shaders is only experimental at
> this point. In order to enable the tessellation path renderer, we need
> to turn off HW tessellation and let it use the indirect draw modes.
>
> Bug: skia:10419
> Change-Id: Ic979978a331c7ad016907cf42b2562b35caab8ab
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344336
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com
Bug: skia:10419
Change-Id: I0bdf34da134a3a34a5c78b35cc0eb86d6e989bb4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/344796
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Adds a vertex shader that maps a variable-length triangle strip to a
stroke and its preceding join. Adds a new op that generates stroke
instances from a path, bins them by log2 triangle strip length (using
SIMD for the calculations), and renders them with indirect draws.
Bug: skia:10419
Change-Id: I6d52df02cffe97d14827c6d66136957f1859f53b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/339716
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Moves the "BenchmarkTarget" class from tessellation benchmarks into a
mock header where it can be reused by other tests and benchmarks.
Change-Id: I344d9ba3d391ff99e10c4ab238684b0b6ada87d0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/333220
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
There's really no reason for these classes to be separate at this
point. Also extracts a "GrStrokeOp" base class that has the
functionality that will be shared with indirect stroking.
Bug: skia:10419
Change-Id: I960d5e6d64f0814ccb4a3852bc627af2b8082a1f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/331860
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This method finds the locations a cubic needs to be chopped at before
it can be passed to the stroke tessellation shader. It's an integral
part of CPU stroke preparation and therefore extremely perf sensitive.
Bug: skia:10419
Change-Id: Ib23c2583b8cfc78814ce52425f7af2c8b2f8b420
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/330314
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
With the addition of vertex data to the thread safe cache we also have
to handle the case where a given SkPath becomes inaccessible and
proactively invalidate the matching entry.
Bug: 1108408
Change-Id: Id11ce2aa10517f7c0772a253634d3c0d13e13460
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/330261
Reviewed-by: Adlai Holler <adlai@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
GrStrokePatchBuilder had the wrong formula. This CL moves the formula to
GrStrokeTessellateOp and calculates it in one spot for the builder and
tessellator both. It also fixes a bug that was hiding behind this
formula error and updates some variable names.
Bug: skia:10419
Change-Id: I908d3960b16cac8dca0d40ff4116a3f5c5beed06
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/328416
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Records programInfos for the stroke ops and for the stencil portions
of the path ops.
We can't prePrepare programInfos for the fill portions yet because it
would require multiple GrPipelines that all reference the same
GrProcessorSet. And GrProcessorSet is currently designed to be
std::moved into one single GrPipeline.
Bug: skia:10419
Change-Id: I3b8c061da181e20d3ff68746cf4b9c61f6d73a88
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/319256
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Adds various optimizations to GrWangsFormula as well as "pow4" variants
of the formula that are quicker than the standard and/or log2 versions.
Uses the pow4 variants in GrStrokePatchBuilder.
Bug: skia:10419
Change-Id: I8478582df5296b088d25808bcaeb93107ff20797
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/318954
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This CL adds a new type GrDstSampleType to say how we will sample the dst.
We add tracking of the GrDstSampleType in the recording of GrOps and
then during execution passing the information along to the GrPipeline.
In general the tracking of GrDstSampleType is a global state of a GrOpsTask
so it is kept separate fro the DstProxyView which is more specific to a
single Op on the GrOpsTask.
Bug: skia:10409
Change-Id: Ie843c31f2e48a887daf96cee99ed159b196cb545
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/315645
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This isn't connected to anything yet and parallels what is done
for the text GrAtlasManager.
The rough pattern is that the GrContext hierarchy implements creating
the atlas but it is only exposed/made available via the flush state
(which derives from GrDrawMeshOp::Target).
TBR=bsalomon@google.com
Bug: 1108408
Change-Id: Iad7b5648877dbd20840e8d47c1d24131098ded94
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/309128
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
SkArenaAlloc has three fields that are used only for reset. Make a
subclass called SkArenaAllocWithReset which has the three
fields, and has the reset functionality.
An example of a reset() that is used instead of using a better scope
is PathOpsAngleAfter in PathOpsAngleTest.cpp.
Change-Id: Ie1965d128dfb7df9e022f4d18460d3f75f33e1a6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307348
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Almost half the files in src/gpu/tessellate were beginning with
"GrTessellat*". This CL reorders the camel case words for more
efficient tab completion and better grouping of related files.
Bug: skia:10419
Change-Id: I3df0ddbb2367cdbd880807a8aef2dd88ec12f830
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/306659
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
