[PATCH] D73796: [mlir][Linalg] Adding support for linalg_matmul with tensors.

[Lei Zhang via Phabricator via llvm-commits]

antiagainst accepted this revision.
antiagainst added a comment.
This revision is now accepted and ready to land.

Neat!



================
Comment at: mlir/test/EDSC/builder-api-test.cpp:1032
+  auto f32Type = FloatType::getF32(&globalContext());
+  auto memrefType = MemRefType::get({-1, -1}, f32Type, {}, 0);
+  auto tensorType = RankedTensorType::get({-1, -1}, f32Type);
----------------
Do we want to use `ShapedType::kDynamicSize` for these magic numbers?


================
Comment at: mlir/test/EDSC/builder-api-test.cpp:1042
+  StructuredIndexed SA(A), SB(B), SC(tensorType);
+  linalg_pointwise_add(SA({i, j}), SB({i, j}), SC({i, j}));
+  linalg_pointwise_max(SA({i, j}), SB({i, j}), SC({i, j}));
----------------
I'm not sure about the conventions here but I'd typically prefer to have small and focused tests so it's easier to maintain and upate.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D73796/new/

https://reviews.llvm.org/D73796