diff options
| author | Max191 <44243577+Max191@users.noreply.github.com> | 2024-11-13 09:49:19 -0500 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-11-13 09:49:19 -0500 |
| commit | 8cc616bc71dfe0648de3843a006ac8827c5fe59d (patch) | |
| tree | 48fe2acfe3db8956e942a8a1e7ec9f9e822985be /llvm/lib/Bitcode/Writer/BitcodeWriter.cpp | |
| parent | 1884ffc41c20b1e08b30eef4e8ebbcc54543a139 (diff) | |
| download | llvm-8cc616bc71dfe0648de3843a006ac8827c5fe59d.zip llvm-8cc616bc71dfe0648de3843a006ac8827c5fe59d.tar.gz llvm-8cc616bc71dfe0648de3843a006ac8827c5fe59d.tar.bz2 | |
[mlir] Clamp UnPackOp tiling sizes from operand tile (#112429)
The `getIterationDomainTileFromOperandTile` implementation for
tensor.unpack did not clamp sizes when the unpack op had extract_slice
semantics. This PR fixes the bug.
The PR also makes a minor change to `tileAndFuseConsumerOfSlice`. When
replacing DPS inits, the iteration domain is needed, and it is computed
from the tiled version of the operation after the initial tiling
transformation. This can result in some extra indexing computation, so
the PR changes it to use the original full sized cloned consumer op.
---------
Signed-off-by: Max Dawkins <max.dawkins@gmail.com>
Diffstat (limited to 'llvm/lib/Bitcode/Writer/BitcodeWriter.cpp')
0 files changed, 0 insertions, 0 deletions