| Commit message (Collapse) | Author | Age |
|---|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Prior to this change, the Slang IR used a single opcode
(`kIROp_Undefined`) to encode all cases of undefined values. The
particular motivation for this change was a need to distinguish those
undefined values that represent a load from an uninitialized memory
location versus other sorts of undefined values. If transforming a
variable into SSA form results in `undefined` values in cases where the
a `load` was executed without a prior `store`, that represents an error
on the programmer's part, and should be diagnosed. However, other cases
of undefined values can arise during program transformation and
optimization, and should not typically result in diagnostics being
emitted.
While it was not the original motivation for this change, it is also
worth noting that the LLVM project has transitioned from initially using
only a single `undef` instruction to having a more nuanced model, and
the same factors that motivated their shift also apply to the Slang IR.
Counter-intuitively, the semantics of undefined values actually need to
be carefully defined.
Concretely, this change splits the pre-existing `undefined` opcode into
two sub-cases:
- `kIROp_LoadFromUninitializedMemory`, to represent the case of loading
from a memory location (such as a local variable) that has not been
initialized.
- `kIROp_Poison`, corresponding to the LLVM `poison` value.
Our poison instruction is intended to have semantics comparable to
LLVM's equivalent. Conceptually, any operation that is invoked with a
poison value as input will (with a few exceptions) produce a poison
value as output. One can think of the behavior of `poison` as similar to
how not-a-number values propagate in floating-point computations: by
default they "infect" the result of any computation they are involved
in. This semantic choice helps to ensure that many optimizations end up
being correct in the presence of undefined values, even if they did not
specifically account for them.
The `kIROp_LoadFromUninitializedMemory` case is comparable to the
combination of `freeze` and `undef` in LLVM. An LLVM `undef` value has
semantics that allow *each* use of that value to be replaced with a
*different* arbitrary value; these semantics cause many optimizations to
only be correct in the absence of undefined values. An LLVM `freeze`
instruction can take an undefined value as input, and produces a single
value that is still arbitrary, but must be consistent across all uses.
The latter semantics are what we want, since a given `load` from an
uninitialized memory location will yield an arbitrary-but-fixed value.
Note that we intentionally do not have a direct analogue to LLVM's
`undef` instruction, because of the way that `undef` causes so many
complications when trying to write optimizations.
We also do not add a `kIROp_Freeze` instruction in this change, but that
is simply because we currently have no need for it.
Existing code that was creating `IRUndefined` values has been updated to
create either `IRPoison` or `IRLoadFromUninitializedMemory` values, as
appropriate to the use case. Code that was checking for the
`kIROp_Undefined` opcode has been updated to either check for both of
the new opcodes (in the case of `switch` statements), or to use
`as<IRUndefined>` to perform a dynamic cast to the common base type of
the two new instructions.
Note that this change does not alter the way that instructions
representing undefined values are typically emitted as ordinary
instructions in the block that produces an undefined value. While
emitting `IRLoadFromUninitializedMemory` as an ordinary instruction is
exactly what we want, the `IRPoison` case would actually be better
represented in Slang IR as a "hoistable" instruction, so that there
would only be a singular `poison` value of each type. Changing
`IRPoison` to be hoistable would be a good follow-up change, but might
run into more challenges depending on what assumptions (if any) the
codebase is making about where undefined values get emitted.
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Close #8572.
The root cause of the issue is that in `_replaceInstUsesWith` call,
if the use of the inst is a generic parameter, and the inst is the data
type of that generic parameter, we could end up of moving the data type
before the generic parameter. This will break the layout of generic
parameters, where all the generic parameters should be laid consecutively
at the beginning of the first block of the generic.
Therefore, we don't make that relocation for such case.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(#8547)
This allows us to specialize functions whose argument is a sub element
of a constant buffer, instead of being only applicable to entire buffer
element. Closes #8421.
This change also implements a proper heuristic to determine when to
specialize the calls and defer the buffer loads.
This PR addresses a pathological case exposed in
`slangpy\slangpy\benchmarks\test_benchmark_tensor.py`, which used to
take 27ms to finish, and now takes 1.25ms.
For example, given:
```
struct Bottom
{
float bigArray[1024];
[mutating]
void setVal(int index, float value) { bigArray[index] = value; }
}
struct Root
{
Bottom top[2];
[mutating]
void setTopVal(int x, int y, float value)
{
top[x].setVal(y, value);
}
}
RWStructuredBuffer<Root> sb;
[shader("compute")]
[numthreads(1, 1, 1)]
void compute_main(uint3 tid: SV_DispatchThreadID)
{
sb[0].setTopVal(1, 2, 100.0f);
}
```
We are now able to specialize the call to `setTopVal` into:
```
void compute_main(uint3 tid: SV_DispatchThreadID)
{
setTopVal_specialized(0, 1, 2, 100.0f);
}
void setTopVal_specialized(int sbIdx, int x, int y, float value)
{
Bottom_setVal_specialized(sbIdx, x, y, value);
}
void Bottom_setVal_specialized(int sbIdx, int x, int y, float value)
{
sb[sbIdx].top[x].bigArray[y] = value;
}
```
And get rid of all unnecessary loads. Achieving this requires a
combination of function call specialization and buffer-load-defer pass.
The buffer-load-defer pass has been completely rewritten to be more
correct and avoid introducing redundant loads.
This PR also adds tests to make sure pointers, bindless handles, and
loads from structured buffer or constant buffers works as expected.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
packing/unpacking. (#8526)
Part of the effort to improve the performance of generated SPIRV code.
The existing lower-buffer-element-type pass works by loading the entire
buffer element content from memory, and translate it to logical type
stored in a local variable at the earliest reference of a buffer handle.
This means that is can generate inefficient code that reads more than
necessary.
Consider this example:
```
struct BigStruct { bool values[1024]; }
ConstantBuffer<BigStruct> cb;
void test(BigStruct v)
{
if (v.values[0]) { printf("ok"); }
}
[numthreads(1,1,1)]
void computeMain()
{
test(cb);
}
```
In IR, the `computeMain` function before lower-buffer-element-type pass
is something like following:
```
func test:
%v = param : BigStruct
%barr = fieldExtract(%v, "values")
%element = elementExtract(%barr, 0)
... // uses %element
func computeMain:
%v = load(cb)
call %test %v
```
The existing lower-buffer-element-type pass will rewrite the bool array
in `BigStruct` into `int` array so it is legal in SPIRV. However, it
does so by inserting the translation on the first `load` of the constant
buffer:
```
struct BigStruct_std430 {
int values[1024];
}
var cb : ConstantBuffer<BigStruct_std430>;
func computeMain:
%tmpVar : var<BigStruct>
call %unpackStorage(%tmpVar, cb)
%v : BigStruct = load %tmpVar
call %test %v
```
This means that the entire array will be loaded and translated to int,
before calling `test`, which only uses one element. It turns out that
the downstream compiler isn't always able to optimize out this
inefficient translation/copy.
This PR completely rewrites the way buffer-element-type lowering is
handled to avoid producing this inefficient code. It works in two parts:
first we turn on the `transformParamsToConstRef` pass for SPIRV target
as well, so we will translate the `test` function to take the `v`
parameter as `constref`. The second part is a redesigned
buffer-element-type pass that defers the storage-type to logical-type
translation until a value is actually used by a `load` instruction.
In this example, after `transformParamsToConstRef`, the IR is:
```
func test:
%v = param : ConstRef<BigStruct>
%barr = fieldAddr(%v, "values")
%elementPtr = elementAddr(%barr, 0)
%element = load(%elementPtr)
... // uses %element
func computeMain:
call %test %cb
```
The new `buffer-element-type-lowering` pass will take this IR, and
insert translation at latest possible time across the entire call graph,
and translate the IR into:
```
func test:
%v = param : ConstRef<BigStruct_std430>
%barr = fieldAddr(%v, "values")
%elementPtr : ptr<int> = elementAddr(%barr, 0)
%element_int = load(%elementPtr)
%element = cast(%element_int) : %bool
... // uses %element
func computeMain:
call %test %cb
```
In this new IR, there is no longer a load and conversion of the entire
array.
See new comment in `slang-ir-lower-buffer-element-type.cpp` for more
details of how the pass works.
This PR also address many other issues surfaced by turning on
`transformParamsToConstRef` pass on SPIRV backend.
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
| |
Fixes: #8018
Changes:
* Do not emit true for `shouldEmitSPIRVDirectly` with a GLSL target
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix duplicate mangled names for interface requirements
Remove linkage decorations from interface method requirement values to prevent
duplicate mangled names and allow DCE to clean up unused functions.
Interface requirements only need the type information, not the linkage,
so the generated IRFunc with export decorations was causing conflicts.
Fixes #7761
Co-authored-by: Yong He <csyonghe@users.noreply.github.com>
* Move linkage decoration removal into default case of switch
Move the removeLinkageDecorations call into the default case of the
switch statement so it only applies to general requirements and not
special cases like associated types.
Co-authored-by: Yong He <csyonghe@users.noreply.github.com>
* Remove conditional check for linkage decoration removal
Apply removeLinkageDecorations unconditionally to interface requirement values
to ensure all requirement values have their linkage decorations removed.
Co-authored-by: Yong He <csyonghe@users.noreply.github.com>
* Change removeLinkageDecorations parameter type from IRGlobalValueWithCode* to IRInst*
This fixes build errors by allowing the function to accept the broader IRInst* type,
making it compatible with the unconditional call in the interface requirement processing.
Co-authored-by: Yong He <csyonghe@users.noreply.github.com>
* format code (#7787)
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Yong He <csyonghe@users.noreply.github.com>
Co-authored-by: slangbot <ellieh+slangbot@nvidia.com>
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Initial plan
* Implement enum-based KnownBuiltin system to replace string comparisons
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Add test for enum-based KnownBuiltin system and verify functionality
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Implement enum-based KnownBuiltin system with direct integer values
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Fix IntVal access and update tests for new enum-based KnownBuiltin system
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Replace hardcoded KnownBuiltin integers with preprocessor enum syntax
- Updated all KnownBuiltin attributes to use $( (int)KnownBuiltinDeclName::EnumValue) syntax
- Added space between parentheses to avoid preprocessor bug: $( (int) instead of $((int)
- Updated both core.meta.slang and hlsl.meta.slang files
- Eliminates preprocessor-time integer conversion, baking enum values directly into meta files
- Maintains same functionality while using type-safe enum references
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Fix IDifferentiablePtr KnownBuiltin mapping regression
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Remove unused IDifferentiablePtrType enum case from KnownBuiltinDeclName
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Clean up temporary AST dump files from testing
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
* Replace hardcoded integer with descriptive constant in KnownBuiltin test
Replace the hardcoded [KnownBuiltin(0)] with a descriptive named constant
GEOMETRY_STREAM_APPEND_BUILTIN to improve code readability and maintainability.
The test now clearly indicates which builtin enum value is being tested.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: Gangzheng Tong <gtong-nv@users.noreply.github.com>
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: csyonghe <2652293+csyonghe@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Gangzheng Tong <gtong-nv@users.noreply.github.com>
Co-authored-by: Gangzheng Tong <tonggangzheng@gmail.com>
|
| |
|
|
|
|
|
|
|
| |
* WIP opaque type decoration fix
* Clearer intent
* Formatting
* Added test for fix
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix for OpUConvert outputting scalar type for mixed sign vector type conversions
* Fix compiler warning
* Added regression test for UConvert vector fix
* Formatting
* getUnsignedType helper
* Formatting
* Fix for addtional int types
* Helper function to convert signed type to unsigned type
|
| |
|
|
|
|
|
|
|
| |
* Cast if there is a signedness mismatch on the swizzle
* Move isSignedType to slang-util and add test
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Handle pointer types when getting type cast style
Closes https://github.com/shader-slang/slang/issues/6025
* Move vertex shader out parameters to return type for Metal
Closes https://github.com/shader-slang/slang/issues/6025
* More asserts
* Make struct instead of tuple
* More layout preservation
* Handle same function result
* more layout
* remove layout
* a
* more debug code
* more debug code
* a
* layout working
* refactored
* more tests
* more tests
* fuse loops
* remove unused comments
* Correct filecheck usage
* debug code
* correct name and order of filecheck vars
* simplify
* Address review comments
fix warning
* simplify handling of simple vertex shaders
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix SPIRV specialization constant with floating-point operations
* Improve test
* WIP
* Restrict `OpSpecConstantOp` allowed operations based on SPIRV specifications
* Fix typo on floating type check
* Emit error on float to int spec cosnt int val casts
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Close #6840.
This PR add supports to use specialize constant in generic parameter, and that parameter can also be used as array size, e.g. following code should work:
```
struct MyStruct<let N: int> { float buffer[N]; }
MyStruct<SpecConstVar> s;
```
- Loose the restriction from Link-Time to SpecializationConstant when extract generic argument
- Tweak the logic of how we decide whether a inst is hoistable. Besides checking existing hoistable flag of each
IRInst, when we detect a IRInst's type is SpecConstRateType, we will treat that inst hoistable. Because IRInst in
global scope can be deduplicated, and every SpecConstRateType inst should be in the global scope or IRGeneric
scope (which will be at global scope after specialization).
- Remove the SpecConstIntVal to IRInst map in IR lowering logic, because we already have way to deduplicate the
global scope IR.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Close #6859
Goal of this PR
We want to support an array whose size can be specialization constant for shared/global variable e.g.
layout (constant_id = 0) const uint BLOCK_SIZE = 64;
shared float buf_a[(BLOCK_SIZE + 5) * 4];
Overview of the solution:
During IndexExpr check, we will loose the restriction to allow SpecConst passing, but the size parameter will not be a constant value because it cannot be folded into a constant, so we will make it follow the same logic as generic parameter value, and the size will be represented by FuncCallIntVal/PolynomialIntVal/DeclRefIntVal.
During IR lowering, we will detect whether there is spec constant in the IntVal, and wrap the IRInst with a SpecConstRateType, and propagate the type though the lowering logic, such that the IntVal representing the array size will have SpecConstRateType.
During spirv emit stage, if we detect that a IRInst has SpecConstRateType, we will emit it as SpecConstantOp.
We have to implement new logic to emit OpSpecConstantOp, the existing emit logic doesn't support emitting OpSpecConstantOp, especially this op can embed arithmetic operation at global scope, where we can only emit arithmetic instruct at local. But there are only few instructs we need to support.
Overview of the solution:
This PR doesn't support generic, and we will create a separate PR to extend that, tracked in #6840.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add Slang Byte Code generation and interpreter.
* Fix compile issues.
* format code
* More compile fix.
* Fix clang issue.
* Fix more clang issues.
* Another clang fix.
* Fix clang issues.
* Fix another clang issue.
* Fix wasm build.
* Update building.md
* Fix test-server.
* Fix compile error.
* Fix bug.
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* initial wip for spirv
* working tiled example
* clean up store and load
* minor fixes
* fix loadAny name
* add initial tests, including broken/unimplemented intrinsics
* fix subscript
* run tests at 16x16, remove not supported arithmetic tests
* minor fixups on implementation
* rename CoopMatMatrixUse
* Update tests to pass validation layers locally
* Add mat-mul-add test and minor fixes
* Add more tests
* Remove dead code
* Add coopMatLoad function and tests, enforce constexpr for matrix layout
* Use getVectorOrCoopMatrixElementType in place of getVectorElementType
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Allow partial specialization of existential arguments.
* Fix.
* Add test case for improved diagnostics.
* Fix compile error.
* Fix tests.
* Fix.
* Fix test.
* Fix compile issue.
* Fix typo.
* Address comment.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix `UseGraph::isTrivial()` test.
* Fix.
* Fix.
* Refactor `UseGraph` and `UseChain`
* Update slang-ir-autodiff-primal-hoist.cpp
* Update all auto-diff locations that handle pointers to treat user pointers as regular values
* Update test to use direct-SPIRV only
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
| |
|
|
|
|
|
| |
Improve performance when compiling small shaders.
Avoid copying witness table entries that are not getting used during linking.
Avoid copying auto-diff related decorations and derivative functions during linking, if the user modules doesn't use autodiff.
Cache operator overload resolution results on global session, so each new Session doesn't need to repetitively run through overload resolution from scratch.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Allow using specialization constants in numthreads attribute
* Add support for GLSL local_size_x_id syntax
* Fix overeager specialization constant parsing
* Add diagnostics for specialization constant numthreads
* Remove unused variable
* Fix local_size_x_id not finding existing specialization constant
* Allow materializeGetWorkGroupSize to reference specialization constants
* Use SpvOpExecutionModeId for modes that require it
* Cleanup specialization constant numthreads code
* Add tests for specialization constant work group sizes
* Fix implicit Slang::Int -> int32_t cast
* Fix querying thread group size in reflection API
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Check whether array element is fully specialized
close #5776
When we start specialize a "specialize" IR, we should
make sure all the elements are fully specialized, but
we miss checking the elements of an array. This change
will check the it.
* add test
* add all wrapper types into the check
* add utility function to check if the type is wrapper type
---------
Co-authored-by: zhangkai <zhangkai@zhangkais-MacBook-Pro.local>
Co-authored-by: Yong He <yonghe@outlook.com>
|
| |
|
|
|
|
|
| |
regression. (#5508)
* Fix IntVal unification logic to insert type casts.
* Fix regression.
|
| |
|
|
|
|
|
|
|
| |
* Move switch statement bodies to their own lines
* format
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
| |
|
|
|
|
|
| |
* format
* Minor test fixes
* enable checking cpp format in ci
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Support mixture of precompiled and non-precompiled modules
This changes the implementation of precompile DXIL modules to
accept combinations of modules with precompiled DXIL, ones without,
and ones with a mixture of precompiled DXIL and Slang IR.
During precompilation, module IR is analyzed to find public functions
which appear to be capable of being compiled as HLSL, and those
functions are given a HLSLExport decoration, ensuring they are emitted
as HLSL and preserved in the precompiled DXIL blob. The IR for those
functions is then tagged with a new decoration AvailableInDXIL, which
marks that their implementation is present in the embedded DXIL blob.
The DXIL blob is attached to the IR as before, inside a EmbeddedDXIL
BlobLit instruction.
The logic that determines whether or not functions should be
precompiled to DXIL is a placeholder at this point, returning true
always. A subsequent change will add selection criteria.
During module linking, the full module IR is available, as well
as the optional EmbeddedDXIL blob. The IR for functions implemented
by the blob are tagged with AvailableInDXIL in the module IR.
After linking the IR for all modules to program level IR, the IR for
the functions marked AvailableInDXIL are deleted from the linked IR,
prior to emitting HLSL and compiling linking the result.
This change also changes the point of time when the module IR is
checked for EmbeddedDXIL blobs. Instead of happening at load time
as before, it happens during immediately before final linking, meaning
that the blob does not need to be independently stored with the module
separate from the IR as was done previously.
Work on #4792
* Clean up debug prints
* Call isSimpleHLSLDataType stub
* Address feedback on precompiled dxil support
Allow for IR filtering both before and after linking.
Only mark AvailableInDXIL those functions which pass
both filtering stages. Functions are corrlated using
mangled function names.
Rather than delete functions entirely when linking with
libraries that include precompiled DXIL, instead convert
the IR function definitions to declarations by gutting
them, removing child blocks.
* Use artifact metadata and name list instead of linkedir hack
* Use String instead of UnownedStringSlice
* Update tests
* Renaming
* Minor edits
* Don't fully remove functions post-link
* Unexport before collecting metadata
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix the issue in emitFloatCast
In emitFloatCast function, we only considered the input type
is float scalar or float vector, so if the input type is a float
matrix type, it will crash.
We should also handle the float matrix type.
Also, we add some diagnose info to point out the source location
where there is error happened, so in the future it's easier to tell
us what happens.
* Add a unit test
* Disable the test for metal
Metal doesn't support 'double'.
"
metal 32023.35: /tmp/unknown-YgHAsJ.metal(15): error : 'double' is not supported in Metal
matrix<double,int(3),int(4)> b_0 = matrix<double,int(3),int(4)> (a_0);
"
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add additional `ImageSubscript` features:
1. Added ImageSubscript support for Metal & a test case
* Merge GLSL/SPIRV/Metal `ImageSubscript` legalization pass
2. Added multisample support to glsl/spirv/metal for when using ImageSubscript
* Added in this PR since the overhaul of the code merges together GLSL/SPIRV/Metal implementation
3. Fixed minor metal texture `Load`/`Read` bugs
* [HLSL methods of access do not support subscript accessor for texture cube array](https://learn.microsoft.com/en-us/windows/win32/direct3dhlsl/texturecubearray)
* removed swizzling of uint/int/float
* other odd bugs which were causing compile errors
note: Compute tests do not work due to what seems to be the GFX backend (causes crash without error report). The tests are disabled.
* disable LOD with texture 1d
seems that LOD for 1d textures need to be a compile time constant as per an error metal throws
* syntax error in hlsl.meta
* static_assert alone with intrinsic_asm error
provides cleaner errors
Note: `static_assert` seems to be unstable and not be fully respected (still require `intrinsic_asm` to avoid a stdlib compile error)
* change comment to `// lod is not supported for 1D texture
* add `static_assert` in related code gen paths
* address review
* address review
* add asserts as per review comment, NOTE: unclear if these should be release 'asserts' as well
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* solve issue by making 'non SV for SPIRV' system semantics, no longer an SV once diagnosed by slang
Problem:
1. SV_InstanceID (HLSL) is allowed as an output semantic with HLSL, it is also allowed for input into pixel shader. We need to account for multiple entry points in 1 file using both of these scenarios at once.
2. SPIRV does not allow `SV_InstanceID` as a builtin, `SV_InstanceID` must be passed as a regular `flat` `Input` from vertex shader.
Solution: Slang needs to treat these SV objects as non built-in's. As a result:
1. Slang needs to allocate for vertex output and fragment Input binding slots for all SV_InstanceID uses (if the target is SPIRV). This allows Slang to prepare an open slot to bind these SV objects to.
2. Slang needs to not emit built in modifier for these not actual built in variables (under the specific circumstances described).
note: `VarLayout` was made not `HOISTABLE` since I don't believe it needs to be `HOISTABLE`.
* The code can be made to work even if `VarLayout` is `HOISTABLE`.
* fix compile warnings
* address review comments and reimplement operand removal
1. remove an operand by selectively copying operands
2. test to ensure `Flat` is in the test shaders
* we do not need to manually add `Flat` in the code since this is done for us in emit-spirv. This is the behavior since `uint` on a fragment `Input` must be `Flat`, else it is a VK validation error.
* address review
remove clone logic
* address review
remove unused function
reserve instead of setCount with ShortList
|
| |
|
|
|
|
|
|
|
| |
* Fix compile failures when using debug symbol.
* Various fixes.
* Fix intrinsic.
* Fix test.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Support derivative functions in compute & capabilities adjustments
fixes #4000
PR implements derivative functions in compute shaders properly so we have the functionality for SPIR-V & GLSL. Tests reflect fragment and compute paths.
PR also adjusts capabilities to correct wrong SPRI-V target capabilities for when using textures.
Remarks:
1. __requireComputeDerivative(); is a intrinsic_op and not modifier since inlining will destroy the modifier.
2. Derivative mode is tied to an entry point decoration `[DerivativeGroupQuad]`/`[DerivativeGroupLinear]` or GLSL syntax ``derivative_group_linearNV`. Default is to set the mode to `[DerivativeGroupQuad]`
* remove -emit-spirv-directly
* fixes
1. fix minor issue fwidth change where I returned the wrong type
2. fix issue where glslang{glsl->spirv} is wrong, so we don't run that test and just run the glsl test & direct spir-v test for intrinsic-texture.slang
* adjust as per review and refine code
1. add test to ensure multi-diverging-in-logic entry points work -- 2 functions which may cause computeDerivatives + 1 that uses, 1 that does not.
2. naming
3. use entry point ref graph for c-like-targets
4. reordered some code to util's and removed `static linline` since that was just for ease of coding on my end (should not have been pushed).
* Grammer
* split up source file + issolate GLSL emit path change.
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
| |
|
|
|
|
|
|
|
| |
* Switch to direct-to-spirv backend as default.
* Fix slang-test.
* Fix.
* Fix.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes #387676* ForceInline SampleLevel to allow decorations to apply
* explictly add all the SPIRVAsmOperand Insts in non-differentiable list, which might get inadvertently processed when these functions are inlined into the main shader
* Support NonUniformResourceIndex for SPIR-V target
Fixes #3876
* add a new IR instruction for NonUniformResourceIndex
* slang ir emitter for nonuniform resource index
* update the hlsl meta slang
* Add test cases for NonUniformResourceIndex access for buffers and textures, with/without cast, nested access etc.
* add default c-like emitter for nonuniformresourceinfo
* added hlsl emitter
* added glsl emitter
* requisites for spirv enabling
- new decorator for nonuniformresourceindex
- emitter for nonuniformresourceindex signature change
* add hasResourceType checker
* add rwStructBuffType in resourcetype checker
* add a case for nonuniformres in emitDecorations
* DO NOT COMMIT: This change adds special handling for RWStructBuf within the isResourceType function, if it is a pointer to this resource, return true to make it work with nonuniformres test
* spirv emitter for decorations - update the emitLocalInst to perform decorations at the end
* added main spirv emitter code
* slang emit spirv bugfix
* hacky way of supporting Call Inst
* move code to cleanup nonuniform inst into helper function
* remove stale codefrom test
* add spirv decoration for nonuniform
* update test to remove global variables
* update coherent-2 test
* update comment for special handling
* update the spirv legalize to handle nested nonuniforms
improved logic that handles call ops, rwstructbuf, nested nonuniforms
etc.
* update nonuniform-array-of-tex test
* missed removing nonuniform inst causing duplicate decorations
* add glsl and hlsl variants of nonuniform tests
* repurpose the hasResource function into something specific for nonuniform inst decoration helper
* clean up comments and code around spirv-legalization to emit nonuniform inst by recursively looking into the inst
* use the helper canDecorateNonUniformInst to convert `nonUniformResourceInfo` inst to decoration
* converted compute/unbounded-array-of-array cross compile test into a simple check test
* update contains Resource helper function to be more generic
* clean up the case for opcall handling with nonuniform resource inst
* update ptr to struct buffer check to be more explicit and rename the function to check for ptr to resource type
* update comments and fix the test for coherent
* fix typos
* update logic on spirv legalize to delete dead instructions - for some reason this doesn't automatically happen
* add comments to declarations
* add NonuniformResourceIndex to the non-differential inst list
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix #3780.
* Fixers #3781.
* Add test for #3781.
* Diagnose error on unsupported builtin intrinsic types.
* Add check for recursion.
* Fix.
* Fix.
* Fix recursion detection.
* Fix.
* Fix.
* Fix recursion logic.
* More fix.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix crash when generating debug info for geometry shaders.
* Fix.
* Fix source language field in DebugCompilationUnit.
* Fix.
* Emit DebugEntryPoint inst.
* Add trivial test.
* Cleanup.
* More cleanup.
|
| |
|
|
|
|
|
| |
* [SPIRV] Add NonSemanticDebugInfo for step-through debugging.
* Fix.
* Fix.
|
| |
|
|
|
|
|
|
|
|
|
| |
* Support pointers in SPIRV.
* Fix test.
* Enhance test.
* Fix test.
* Cleanup.
|
| |
|
|
|
| |
parent. Previously special case was added to handle IRDecoration similarly. Replace this with a common method getBlock that traverses the parent chain till it gets to the Block (#3486)
Fixes bug #3432
|
| |
|
|
|
|
|
|
|
| |
* Fix crash when writing to `no_diff` out parameter.
* Fix.
---------
Co-authored-by: Yong He <yhe@nvidia.com>
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* SPIRV compiler performance fixes.
* Cleanup.
* update project files
* Cleanup debug code.
* Make redundancy removal non-recursive.
---------
Co-authored-by: Yong He <yhe@nvidia.com>
|
| |
|
|
|
|
|
|
|
| |
* More direct-SPIRV fixes.
* Fix array-reg-to-mem.
---------
Co-authored-by: Yong He <yhe@nvidia.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Support `constref` parameters passing.
* Fix.
* Fix.
* Add test and diagnostic on mix use of __constref and no_diff.
* check for [constref] on differentiable member method.
---------
Co-authored-by: Yong He <yhe@nvidia.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Various SPIRV fixes.
- Geometry shader support (WIP).
- Fix texture get dimension and load.
- Fold global GetElement(MakeArray/MakeVector) insts.
- Call spvopt to inline all functions.
- Translate OpImageSubscript.
- Emit struct member names and global variable names.
- Fix lowering of OpBitNot -> OpNot, instead of OpBitReverse.
* Fix test.
* Fix geometry shader.
* Fix geometry shader emit.
* Add atomic Image access test.
* Fix tests.
* don't fail if spirv-opt fails.
* Update comments.
* Fix test.
* Cleanups.
* indentation
---------
Co-authored-by: Yong He <yhe@nvidia.com>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix issue with trivial loop detection
* Fix issue with unreachable blocks in break elimination
Add logic to avoid eliminating loops with multi-level breaks.
* Incorporate feedback
- Use a boolean for multi-level break check
- Use dominator trees for region check instead of exhaustive enumeration
- Fix potential issue with enumerating parent break blocks.
* fix
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Misc. SPIRV Fixes, Part 2.
* Fix up.
* Fix.
* Add system smenatic values.
* 16 bit int and floats, matrix/vector reshape, bool ops.
* Fix.
* Fix.
* Allow push constant entry point params.
* entrypoint params.
* swizzleSet and swizzledStore.
* packoffset.
* string hash.
* Fix.
* Matrix arithmetics.
---------
Co-authored-by: Yong He <yhe@nvidia.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Generalize collectInductionValues
* Support affine transformations of loop index as induction variables
* Test for generalized induction value collection
* Neaten inductive variable finding
* Make types more specific
* Add loop inversion pass
* Test output changes after loop inversion
* Store the type of implication success when finding inductive variables
* Test that loop induction finding does not alway succeed
* Support chains of additions and branches of additions in induction variable finding
* Use c++17 for downstream compilers
* Wiggle expected output for cross compile test after loop inversion
* Add loop inversion test
* Simplify IfElse instructions with a single trivial block
* Invert loops with a user inserted break
* Limit loop inversion to loops with a 4 instruction or less comparison block
* regenerate vs projects
|
