Magic-Mirror/llvm-capstone
1
0
Fork 0
mirror of https://github.com/capstone-engine/llvm-capstone.git synced 2024-11-30 09:01:19 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity
auto-sync
llvm-capstone/mlir/test/python/dialects
History
Guray Ozen 12c241b365
[MLIR][NVVM] Explicit Data Type for Output in wgmma.mma_async (#78713) 
The current implementation of `nvvm.wgmma.mma_async` Op deduces the data
type of the output matrix from the data type of struct member, which can be
non-intuitive, especially in cases where types like `2xf16` are packed
into `i32`.

This PR addresses this issue by improving the Op to include an explicit
data type for the output matrix.

The modified Op now includes an explicit data type for Matrix-D (<f16>),
and looks as follows:

```
%result = llvm.mlir.undef : !llvm.struct<(struct<(i32, i32, ...
nvvm.wgmma.mma_async
    %descA, %descB, %result,
    #nvvm.shape<m = 64, n = 32, k = 16>,
    D [<f16>, #nvvm.wgmma_scale_out<zero>],
    A [<f16>, #nvvm.wgmma_scale_in<neg>, <col>],
    B [<f16>, #nvvm.wgmma_scale_in<neg>, <col>]
```
2024-01-22 08:37:20 +01:00
..
gpu [MLIR][LLVM] Remove last typed pointer remnants from tests (#71232) 2023-11-04 14:13:31 +01:00
linalg [mlir][linalg] regionBuilder for transpose, broadcast (#69742) 2023-10-20 16:14:46 -05:00
sparse_tensor [mlir][sparse] rename DimLevelType to LevelType (#73561) 2023-11-27 14:27:52 -08:00
affine.py [mlir][python] fix affine test 2023-12-07 16:21:57 -06:00
amdgpu.py [MLIR][python bindings] add vendor gpu dialects 2023-08-13 16:45:20 -05:00
arith_dialect.py [mlir][python] meta region_op (#75673) 2023-12-21 11:20:29 -06:00
arith_llvm.py [mlir][arith] Add overflow flags support to arith ops (#78376) 2024-01-17 06:12:23 +03:00
async_dialect.py
builtin.py [mlir][python] Remove __str__ from bindings of StringAttr. 2023-09-01 07:35:54 +00:00
cf.py [mlir][CAPI, python bindings] Expose Operation::setSuccessor (#67922) 2023-10-02 15:37:25 -05:00
complex_dialect.py
func.py [mlir][python] simplify extensions (#69642) 2023-10-19 18:07:06 -05:00
llvm.py [mlir][python bindings] generate all the enums 2023-08-23 15:03:55 -05:00
math_dialect.py
memref.py [mlir][python] add MemRefTypeAttr attr builder (#76371) 2024-01-06 16:42:14 -06:00
ml_program.py
nvgpu.py [MLIR][python bindings] add vendor gpu dialects 2023-08-13 16:45:20 -05:00
nvvm.py [MLIR][NVVM] Explicit Data Type for Output in wgmma.mma_async (#78713) 2024-01-22 08:37:20 +01:00
ods_helpers.py Finish renaming getOperandSegmentSizeAttr() from operand_segment_sizes to operandSegmentSizes 2023-08-09 19:37:01 -07:00
openmp_ops.py [mlir][python bindings] turn on openmp 2023-08-23 18:17:04 -05:00
pdl_ops.py
pdl_types.py
python_test.py [mlir][python] enable registering dialects with the default Context (#72488) 2023-11-27 19:26:05 -06:00
quant.py
rocdl.py [MLIR][python bindings] add vendor gpu dialects 2023-08-13 16:45:20 -05:00
scf.py [mlir][python] fix scf.for_ convenience builder (#72170) 2023-11-13 20:25:41 -06:00
shape.py
spirv_dialect.py [mlir][spirv] Add support for C-API/python binding to SPIR-V dialect (#76055) 2024-01-02 08:11:44 -08:00
tensor.py [mlir][python] meta region_op (#75673) 2023-12-21 11:20:29 -06:00
tosa.py
transform_bufferization_ext.py [mlir][bufferization] Remove allow-return-allocs and create-deallocs pass options, remove bufferization.escape attribute (#66619) 2023-09-18 16:44:48 +02:00
transform_extras.py [MLIR][transform][python] Introduce abstractions for handles to values and parameters (#77305) 2024-01-15 10:31:22 +01:00
transform_gpu_ext.py [mlir][gpu][transform] Provide better error messages and avoid crashing in MapForallToBlocks. 2023-09-04 14:11:38 +00:00
transform_loop_ext.py [mlir][SCF] Use transform.get_parent_op instead of transform.loop.get_parent_for (#70757) 2023-10-31 18:36:40 +09:00
transform_memref_ext.py [mlir][memref][transform] Add new alloca_to_global op. (#66511) 2023-09-21 18:17:00 +02:00
transform_nvgpu_ext.py [mlir] enable python bindings for nvgpu transforms (#68088) 2023-10-03 14:52:52 +02:00
transform_sparse_tensor_ext.py [mlir][sparse] Generates python bindings for SparseTensorTransformOps. (#66937) 2023-09-20 15:35:50 -07:00
transform_structured_ext.py [mlir][ODS] Omit printing default-valued attributes in oilists (#68880) 2023-10-13 10:22:05 +01:00
transform_tensor_ext.py [mlir][python bindings] generate all the enums 2023-08-23 15:03:55 -05:00
transform_vector_ext.py [mlir][ODS] Omit printing default-valued attributes in oilists (#68880) 2023-10-13 10:22:05 +01:00
transform.py [mlir][python] Add support for arg_attrs and other attrs to NamedSequenceOp 2023-11-08 13:42:16 +00:00
vector.py [mlir][python bindings] generate all the enums 2023-08-23 15:03:55 -05:00