diff options
| author | maxbartel <bartel@roofline.ai> | 2025-10-08 14:39:23 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-10-08 13:39:23 +0100 |
| commit | f53b6249c24005d1a6208cd9e355595eb6519dc0 (patch) | |
| tree | 660d3fa53031240225f3f41cd44b200e191dfc98 /lldb/test/API/lua_api/TestLuaAPI.py | |
| parent | 81589a39bff58e5a750ea7b656f66cf941803e3b (diff) | |
| download | llvm-f53b6249c24005d1a6208cd9e355595eb6519dc0.zip llvm-f53b6249c24005d1a6208cd9e355595eb6519dc0.tar.gz llvm-f53b6249c24005d1a6208cd9e355595eb6519dc0.tar.bz2 | |
[MLIR][Linalg] Fix empty tensor assumptions for linalg.pack decomposition (#160246)
The original code seemed to assume that the tiling dimensions for the
tensor.empty op before applying the transposing were always the last
dimensions. However, pack allows you to choose any dimension to tile.
The easiest way I found to solve this is to prefill the SmallVector with
1s of size (srcRank - numberOfTiles) and then appending the tile sizes.
This way I could also get rid of the first loop in the code.
Diffstat (limited to 'lldb/test/API/lua_api/TestLuaAPI.py')
0 files changed, 0 insertions, 0 deletions