(This mirrors an optimization performed in the constant folder.)
Expressions like `OpIEqual %20 %20` or `OpFUnordNotEqual %15 %15` can be
replaced by `true` or `false` on sight. The GLSL spec makes it clear
that checking for NaN is optional:
4.7.1 Range and Precision
"... NaNs are not required to be generated. Support for signaling NaNs
is not required and exceptions are never raised. Operations and built-in
functions that operate on a NaN are not required to return a NaN as the
result."
Change-Id: I2e29b659a73582e9ade0eb61f70f7d362a007c50
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/531550
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
We now have two functions `writeOpLoad` and `writeOpStore` which are
in charge of writing SpvOpLoad and SpvOpStore instructions.
`writeOpStore` also keeps track of pointer stores in a "store cache."
Subsequent loads from that same pointer will be found in the cache and
will return the value stored in that pointer instead.
Such a cache definitely cannot work in the face of control flow, so we
make the following concessions:
- `pruneReachableOps` is now `pruneConditionalOps`. Any pointers that
are altered inside a potentially-unreachable block are cleared from
the cache entirely.
- The entire store cache is cleared at all OpLabels within a loop.
The cache also cannot work in the presence of swizzled stores, so we
make another significant concession:
- The entire store cache is cleared whenever we store into a non-memory
pointer (e.g., assigning into a swizzled LValue, such as `foo.xz`).
Despite these significant limitations, this manages to dramatically
shrink many real-world examples.
Change-Id: I0981a0cf7b45b064e153e9ada271494c8e00cad5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/530054
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Previously, we would determine the operandType by calling
`getActualType`. This function converts half-precision types to full-
precision ones, which seems to have been unintentional. Fortunately, the
operand type is not actually emitted into the SPIR-V by most code paths
(most paths use the resultType instead) so it was not a significant
impact in practice. A few matrix-based paths emitted ops using this type
and these paths now emit RelaxedPrecision as expected.
Change-Id: I32f4c0327427476fee6b78953284818b7970b6e8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/530543
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Arman Uguray <armansito@google.com>
Previously, we only handled the simple case of extracting from an
OpConstantComposite. Now we also handle the complex case of extracting
from an OpCompositeConstruct, where vectors can be composed of other
vectors.
This is particularly challenging because OpCompositeConstruct can
contain SpvIds from almost any other instruction, so we need to be
able to decode those instructions and figure out their type. For
instance:
%5 = OpFAdd Vec2 %1 %2
%6 = OpFAdd Scalar %3 %4
%7 = OpCompositeConstruct FloatVec3 %5 %6
%8 = OpCompositeExtract %7 2
The %8 (OpCompositeExtract) could be replaced with %6 but we need to
peek at the type in *both* OpFAdd instructions to decode this. It
only works when the affected instructions are in-cache, so many
opportunities are currently not optimized because their code still
uses the original, uncached form of writeInstruction.
Change-Id: I5719ae6284f32e1d6f2c898eca282c22b94fc764
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529743
Reviewed-by: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Bug: skia:13219
Change-Id: I57c5c2aa40e6eb85d5e6045d6f3374d0379efd39
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/530337
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This removes the assert from the SPIR-V generator so the
tests compile. The generated SPIR-V is incorrect. The next
CL fixes the generator, and restores the assert.
Change-Id: I77b507cf7fb5eac481322887000bd1c73cd5c899
Bug: skia:13219
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/530336
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Previously, we stringized the types and put them into fTypeMap. Using
the op cache is a simpler mechanism that should work equally well.
Output diffs are almost all ID reorderings. In a few cases we
managed to deduplicate function types that stringize differently but
come out the same in SPIR-V (e.g. no float/half distinction).
Change-Id: If7de5b2dafa12d05c3c2c497a243e9e3908dfee7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529805
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Change-Id: Ic19d7591c90f75f04dd1c58b406f2b770f9780c7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529351
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
If we start with an OpConstantComposite, then we do an
OpCompositeExtract from it, we can look up the result directly from op
cache and avoid doing any work. This helps our matrix code a lot.
Change-Id: Idfbdc0c69676b9c1e91cdc57bf0d6382b9b5d8d4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529339
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Previously, we didn't usually generate OpConstantComposite ops for
matrices. Now, a matrix assembled from constants should come out as a
constant.
Change-Id: I458718901686dffb84e4079a81017d61195420d3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/529338
Reviewed-by: Arman Uguray <armansito@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
The output changes here are almost entirely a wash, because we already
had support for caching scalars and vectors. Almost all changes are just
inconsequential reorderings of IDs, and the removal of RelaxedPrecision
decorators on constants (which were not meaningful).
Change-Id: I45340c4a240cb504b7c4a934b3db178d2f39ec99
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/528709
Reviewed-by: Arman Uguray <armansito@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
`writeComposite` can write and deduplicate constants, so it's preferable
to manually emitting an OpCompositeConstruct opcode.
Change-Id: I0c4ac8f8a456c8561c0b6a90cd316934f20895e8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/528638
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
`writeComposite` can write and deduplicate constants, so it's preferable
to manually emitting an OpCompositeConstruct opcode.
Change-Id: Ie5c23af76822da762eadac8ff0ab0c6cc0febd31
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/528637
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Some GPUs (Adrenos in particular) perform noticeably better when we
use OpConstantComposite instead of OpCompositeConstruct. This also gives
us some deduplication of redundant ops.
Change-Id: I53b7a3e1cf61e51647a661a08ff4c7b53ee60f10
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/528636
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Previously, when we vectorized scalars, we would not place them in the
constant buffer, so we could emit simple vectors like (0,0,0) or (1,1,1)
more than once. Now, we use `writeConstructorSplat` to vectorize, which
knows how to write constants.
Change-Id: Ic97c0ce5415fd46ff8c7fb7dac9205844633ef3a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/527921
Reviewed-by: Arman Uguray <armansito@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
On modern hardware, this will give the correct result for `NaN != x`
(true).
Change-Id: I9683f74756da5da5f34ccacec02c1f2449791f26
Bug: skia:12977
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/513317
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Arman Uguray <armansito@google.com>
Commit-Queue: Arman Uguray <armansito@google.com>
GPUs that failed continued to fail when I put in error bars like
`distance(a, b) <= 0.001`, so they're just disabled entirely now.
Presumably their results are very busted.
Change-Id: I0f1b80f661563a20630740f8cfb6ef69f2a47934
Bug: skia:11209, skia:12858
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/503817
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Removed a special case from `writeFunctionCallArgument` which avoided
using a scratch variable for out params; now we always use the scratch
variable and copy it back to the original variable at the end.
Change-Id: I0e446a3fde6d19554943384210bd911f6f9c8cfa
Bug: skia:11052, skia:11919, skia:12858
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/501836
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit a97a6769b5.
Reason for revert: breaking bot Housekeeper-PerCommit-CheckGeneratedFiles -- see http://screen/5FN75fF9tvcQFKR
Based on the diff, I think this just needs to be synced up to latest code and a rebuild should fix it.
Original change's description:
> [skslc] Generate .hlsl test output files
>
> - The build now generates HLSL output when `skia_compile_sksl_tests` is
> enabled.
> - The "blend" and "shared" tests have been enabled for HLSL with the
> exception of 6 tests that exercise intrinsic inverse hyperbolic
> functions, which don't have HLSL equivalents.
>
> Bug: skia:12691, skia:12352
> Change-Id: Ia970f878f75ff58e8e3d47249c2dc2f756c165b4
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/482778
> Reviewed-by: John Stiles <johnstiles@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
> Auto-Submit: Arman Uguray <armansito@google.com>
> Commit-Queue: Arman Uguray <armansito@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
Bug: skia:12691, skia:12352
Change-Id: Iaad607d48edd136eee2b60e48c0643b6e90179e9
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/499216
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
- The build now generates HLSL output when `skia_compile_sksl_tests` is
enabled.
- The "blend" and "shared" tests have been enabled for HLSL with the
exception of 6 tests that exercise intrinsic inverse hyperbolic
functions, which don't have HLSL equivalents.
Bug: skia:12691, skia:12352
Change-Id: Ia970f878f75ff58e8e3d47249c2dc2f756c165b4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/482778
Reviewed-by: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Auto-Submit: Arman Uguray <armansito@google.com>
Commit-Queue: Arman Uguray <armansito@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Note that the 2D cross product isn't defined. There are at least two
possible interpretations of what that might mean. This name makes it
clearer that we're asking for the length of the resulting vector, if
we computed the 3D cross product (assuming Z == 0 for both vectors).
It also eliminates name overlap between builtin functions and actual
intrinsics.
Change-Id: I24e8bc0ab2ec91aaace20f0dd3e8565c10bd44a0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/484440
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
We never used it internally, but the shaders used by Filament rely on
it. It doesn't exist in ES2 so this doesn't affect Runtime Effects.
Change-Id: Idb2afb15ff160b950ad02101bf6381a5d5c56468
Bug: skia:12635, skia:11209
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/470156
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
This will hopefully improve performance on lower-end GPUs.
Change-Id: I9c2ee6dc31acd08bec0bfb5f59edc3cf90163f9e
Bug: skia:12339
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/465078
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
This guards against unexpected results when dfdy is used in complex
expressions. In practice, I'm not aware of this causing any trouble.
Change-Id: Ia476e57936969d248273856a94d5c403b47c29b4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/465379
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This guards against unexpected results when dfdy is used in complex
expressions. In practice, I'm not aware of this causing any trouble.
Change-Id: I639bef465d7907049d79681a49f9be67b4c435a6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/465378
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit 9fc189f1cb.
Reason for revert: shader compile failure on AndroidOne-GPU-Mali400MP2 devices
Original change's description:
> Wrap 'u_rtFlip.y * dfdy()' in parentheses.
>
> This guards against unexpected results when dfdy is used in complex
> expressions. In practice, I'm not aware of this causing any trouble.
>
> Change-Id: I58d4762871481fdb4c173b570e4d5d6edf657af7
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/465077
> Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
> Auto-Submit: John Stiles <johnstiles@google.com>
Change-Id: Idfaa9316d657717d5ee7117837c9cc9c3d4ee189
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/465377
Auto-Submit: Greg Daniel <egdaniel@google.com>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
This guards against unexpected results when dfdy is used in complex
expressions. In practice, I'm not aware of this causing any trouble.
Change-Id: I58d4762871481fdb4c173b570e4d5d6edf657af7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/465077
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
This was mistakenly using dFdx in some portions (copy-paste error).
Change-Id: Ifb159b3c44185d9166c10725b24002a28a0895b2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/464381
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
The inverse, outerProduct and matrixCompMult polyfill functions in Metal
were written assuming that all float matrices would use the `float`
type. They now use a template so that `half` matrices will work too.
Change-Id: I7696c8ad1e4aaffbd71c56b9245485e74cd96c5a
Bug: skia:12339
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/463338
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit eb68973c2f.
Reason for revert: ES2 conformance test checks this
Original change's description:
> Disallow matrix ctors which overflow a column.
>
> The GLSL spec allows matrix constructors containing vectors that would
> split between multiple columns of the matrix. However, in practice, this
> does not actually work well on a lot of GPUs!
>
> - "cast not allowed", "internal error":
> Tegra 3
> Quadro P400
> GTX 660
> GTX 960
> - Compiles, but generates wrong result:
> RadeonR9M470X
> RadeonHD7770
>
> Since this isn't a pattern we expect to see in user code, we now report
> it as an error at compile time. mat2(vec4) is treated as an exceptional
> case and still allowed.
>
> Change-Id: Id6925984a2d1ec948aec4defcc790a197a96cf86
> Bug: skia:12443
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/449518
> Commit-Queue: John Stiles <johnstiles@google.com>
> Auto-Submit: John Stiles <johnstiles@google.com>
> Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Bug: skia:12443
Change-Id: I5a32744c88b9b830ad657488824c8c7dd0b0a652
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/458056
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>
`optimize_comparison` asserted that its inputs were numbers. However,
it's also valid to compare boolean inputs. Fortunately, other than the
over-zealous assertion, the actual logic worked fine.
Change-Id: I8a9db000274b4993a4c303efa223a1ed72461a87
Bug: oss-fuzz:39513
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/455296
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
`input` is a reserved word in GLSL. http://screen/85m4iRwvJRadKbV
Change-Id: Iffc0a47d916a2419a27767902c839e09bfa7fe26
Bug: skia:11115
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/454736
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
The GLSL spec allows matrix constructors containing vectors that would
split between multiple columns of the matrix. However, in practice, this
does not actually work well on a lot of GPUs!
- "cast not allowed", "internal error":
Tegra 3
Quadro P400
GTX 660
GTX 960
- Compiles, but generates wrong result:
RadeonR9M470X
RadeonHD7770
Since this isn't a pattern we expect to see in user code, we now report
it as an error at compile time. mat2(vec4) is treated as an exceptional
case and still allowed.
Change-Id: Id6925984a2d1ec948aec4defcc790a197a96cf86
Bug: skia:12443
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/449518
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Not a big deal necessarily, but considering using this logic in GLSL as
well, and I'm less confident that your average GLSL ES driver will
optimize away the separate array loads.
Change-Id: I6a9f0d18c0fac138f64ad6426670f615e17f3492
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/449099
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
We already had a test case here, but it wasn't actually in operation.
The test has been split into ES2 (square) and ES3 (non-square) halves,
returns the color like a proper runtime effect, and it's now running in
dm.
Also, Metal doesn't natively support matrixCompMult, so it injects a
helper function; I tweaked the helper so it no longer requires an extra
result variable.
Change-Id: Ie79242768966fcbe879ad73461d17b4fb8e55670
Bug: skia:12202
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/448117
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Throughout SkSL we've begun using doubles as a convenient way to store
any SkSL value (int, float, bool) in a single type. This idea has now
been extended to literals. Rather than having three expression kinds for
integers, floats and boolean literals, we can have just one. These can
be accessed in a type-specific way (`floatValue`, `intValue`, and
`boolValue` return the expected type, or assert if it's not the
matching type), or in a type-agnostic way (`value` will return a double
and works on any type of Literal).
This allows us to remove a complex template trick (Literal<T> is gone),
removes two redundant Expression types, and and lets us reduce our code
size in ConstantFolder, FunctionCall, etc.
Most of the conversion process was pretty straightforward:
* `IntLiteral::Make` becomes `Literal::MakeInt`
* `x.is<IntLiteral>()` becomes `x.isIntLiteral()`
* `x.as<IntLiteral>.value()` becomes `x.as<Literal>.intValue()`
Change-Id: Ic328533611e4551669c7fc9d7f9c03e34699f3f6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/447836
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: If57fb5acbf5bd0cfeadc54dd12c3ba1da0840491
Bug: skia:12202
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/447416
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Compile-time optimization is not yet implemented so the generated code
contains a lot of checks which will be optimized away in a followup CL.
Change-Id: I83b5df8580a6712686d18812e3848a703feac315
Bug: skia:12202
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/447300
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Metal doesn't natively offer this intrinsic at all, but we now write an
equivalent template function whenever the intrinsic is encountered.
Proper testing will be added in a followup CL (when outerProduct is made
available in public SkSL).
Change-Id: Ie8d6bf8d735d0ab45b7559be68036b08c5802365
Bug: skia:12202
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/447296
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This includes compile-time optimization and tests.
The unit test is disabled in a followup CL
(http://review.skia.org/447057) because it exposes a Radeon 5300M bug
in OpenGL.
Change-Id: I8b2f0411358aeb68c4edfeb0bd7a2814c4be1f40
Bug: skia:12202
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/447056
Reviewed-by: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
It turned out that Metal had equivalent intrinsics/casts all along; we
just needed to emit them.
Tests will be improved in a followup CL which adds the ES3-compatible
packing intrinsics into sksl_public.
Change-Id: Iec8a20b9f9fe9b1badea2944eb0b1f0a17c74560
Bug: skia:12351
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/446744
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
The test has been improved and now covers a variety of values. Because
this intrinsic has side-effects (an out-param), we do not support
optimizing it or treating it as a constant-expression.
The modf documentation doesn't mention anything about constant-
expression support or lack thereof. Experimentally, modf is also not
treated as a constant-expression by Apple GLSL or glslang:
http://screen/4RWwYKr6vCjxCPQhttp://screen/45ttDTVAFGDRyxP
Change-Id: I15bb1de80e90fa97ddf8e9d3803352603b9608d0
Bug: skia:12202
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/446396
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
The code in writeFunctionCall which supported swizzled out-params has
been factored out into a pair of helper functions. The code for emitting
intrinsics now relies on these helpers when emitting out-params.
Change-Id: I4436185ae107d70b529e7e1ea0dd89f844c6a673
Bug: skia:11052
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/446719
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit 0f4304e6e7.
Reason for revert: breaks Adreno 6xx
Original change's description:
> Add RelaxedPrecision decoration to function-call temp vars.
>
> This is really same basic issue as http://review.skia.org/446640. We
> were creating a temp variable but ignoring its type's precision.
>
> Change-Id: I9a5fedd7ada864d36757fc196f42ff95bac7d706
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/446718
> Commit-Queue: John Stiles <johnstiles@google.com>
> Auto-Submit: John Stiles <johnstiles@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
Change-Id: I6ae4e264b60f7f38a1abb5f1d0324461a33c896d
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/446742
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
This is really same basic issue as http://review.skia.org/446640. We
were creating a temp variable but ignoring its type's precision.
Change-Id: I9a5fedd7ada864d36757fc196f42ff95bac7d706
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/446718
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit 9155b338bb.
Reason for revert: disable test for GLSL + Adreno 6xx
Original change's description:
> Revert "Add ES3 intrinsics isinf/isnan to public SkSL ES3."
>
> This reverts commit e43714f490.
>
> Reason for revert: Several Pixel (Adreno) devices failing the test
>
> Original change's description:
> > Add ES3 intrinsics isinf/isnan to public SkSL ES3.
> >
> > The ES3 spec doesn't mandate that `isnan` actually has to do anything,
> > so the Isnan test is not enabled. (It doesn't work on my personal
> > machine unless I make the NaN detectable at compile-time.)
> >
> > We do not support these functions in constant-expressions, as we
> > currently avoid optimizing anything into a non-finite value; we leave
> > expressions alone if we calculate a NaN/inf result for their value.
> >
> > Change-Id: Ibfdfb47b6e6134165c8780db570de04a916d2bfa
> > Bug: skia:12022
> > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/441581
> > Auto-Submit: John Stiles <johnstiles@google.com>
> > Reviewed-by: Brian Osman <brianosman@google.com>
> > Commit-Queue: John Stiles <johnstiles@google.com>
>
> TBR=brianosman@google.com,ethannicholas@google.com,johnstiles@google.com,skcq-be@skia-corp.google.com.iam.gserviceaccount.com
>
> Change-Id: I89899ed391aa870350d0452bab4a0fb75bd7be38
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: skia:12022
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/441716
> Reviewed-by: Brian Osman <brianosman@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
Bug: skia:12022, skia:12377
Change-Id: Ib149dbc1138feb3ee2bf6f7e31e9e8a9414560bc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/441884
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
