Details
Details
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
mlir/lib/Conversion/StandardToLLVM/StandardToLLVM.cpp | ||
---|---|---|
2412 | Can't we just emit a constant instead? |
mlir/test/Conversion/StandardToLLVM/convert-to-llvmir.mlir | ||
---|---|---|
1307 | Irrelevant to this patch, but shouldn't we use i32 for the rank if bitwidth of index is 32? |
mlir/test/Conversion/StandardToLLVM/convert-to-llvmir.mlir | ||
---|---|---|
1301 | Does it change anything? |
mlir/test/Conversion/StandardToLLVM/convert-to-llvmir.mlir | ||
---|---|---|
1307 |
Maybe rephrase into a positive: RankOp should be folded away for ranked operands.