llvm-capstone/mlir
Guray Ozen 763109e346
[mlir][gpu] Use known_block_size to set maxntid for NVVM target (#77301)
Setting thread block size with `maxntid` on the kernel has great
performance benefits. In this way, downstream PTX compiler can do better
register allocation.

MLIR's `gpu.launch` and `gpu.launch_func` already has an attribute
(`known_block_size`) that keeps the thread block size when it is known.
This PR simply uses this attribute to set `maxntid`.
2024-01-08 14:49:19 +01:00
..
benchmark/python [mlir][benchmark] Fix broken benchmark script (#68841) 2023-12-06 12:17:53 +05:30
cmake/modules Revert "[mlir] Consider mlir-linalg-ods-gen as a tablegen tool in build (#75093)" 2024-01-04 02:01:16 +05:00
docs [mlir][gpu] Improve gpu-lower-to-nvvm-pipeline Documentation (#77062) 2024-01-05 12:51:25 +01:00
examples [MLIR][LLVM] Add distinct identifier to DICompileUnit attribute (#77070) 2024-01-08 07:42:33 +01:00
include [mlir] Add explicit call to flush 2024-01-08 08:04:13 +00:00
lib [mlir][gpu] Use known_block_size to set maxntid for NVVM target (#77301) 2024-01-08 14:49:19 +01:00
python [mlir][python] add MemRefTypeAttr attr builder (#76371) 2024-01-06 16:42:14 -06:00
test [mlir][gpu] Use known_block_size to set maxntid for NVVM target (#77301) 2024-01-08 14:49:19 +01:00
tools add prop-dict support for custom directive for mlir-tblgen (#77061) 2024-01-05 12:37:24 +01:00
unittests [MLIR][Presburger] Implement IntegerRelation::mergeAndAlignSymbols (#76736) 2024-01-07 17:06:52 +05:30
utils [mlir][arith] Rename operations: maxfmaximumf, minfminimumf (#65800) 2023-09-11 22:02:19 -07:00
.clang-format
.clang-tidy
CMakeLists.txt [mlir] Add config for PDL (#69927) 2024-01-03 20:37:20 -08:00
LICENSE.TXT
README.md

Multi-Level Intermediate Representation

See https://mlir.llvm.org/ for more information.