llvm-project/mlir/lib/Dialect/Linalg
Aart Bik 8b124c19f5 [mlir][sparse] adjust output shape inference to new tensor abstraction
Nicolas changed the tensor abstraction so that every output has
its own shape definition. This simplifies the "inference" that
was used in the sparse compiler.

Reviewed By: penpornk

Differential Revision: https://reviews.llvm.org/D94119
2021-01-05 15:31:39 -08:00
..
Analysis [mlir][Linalg] Revisit the Linalg on tensors abstraction 2020-12-21 12:29:10 -08:00
EDSC [mlir][Linalg] Revisit the Linalg on tensors abstraction 2020-12-21 12:29:10 -08:00
IR [mlir] Add canonicalization pattern out_tensor->linalg->dim to out_tensor->dim. 2021-01-05 15:15:21 +01:00
Transforms [mlir][sparse] adjust output shape inference to new tensor abstraction 2021-01-05 15:31:39 -08:00
Utils [mlir][Linalg] Revisit the Linalg on tensors abstraction 2020-12-21 12:29:10 -08:00
CMakeLists.txt [mlir][EDSC] Refactor dependencies involving EDSCs. 2020-01-15 09:34:29 -05:00