[PATCH] D74544: [MLIR] Add naive fusion of parallel loops.

[Alexander Belyaev via Phabricator via llvm-commits]

pifon2a marked 9 inline comments as done.
pifon2a added inline comments.


================
Comment at: mlir/lib/Dialect/LoopOps/Transforms/ParallelLoopFusion.cpp:100
+/// in secondPloop.
+static bool haveNoReadWritesToSameBufs(ParallelOp firstPloop,
+                                       ParallelOp secondPloop) {
----------------
herhut wrote:
> Would it not suffice to call `haveNoReadAfterWritesExpectSameIndex` with arguments reversed?
No, because we are using BlockAndValueMapping that has only a map from IVs of ploop1 to IVs to ploop2. We can have the inverse map though.


================
Comment at: mlir/test/Dialect/Loops/parallel-loop-fusion.mlir:261
+
+func @do_not_fuse_unmatching_read_write_patterns(
+    %A: memref<2x2xf32>, %B: memref<2x2xf32>, %C: memref<2x2xf32>) {
----------------
herhut wrote:
> How is this different from above?
renamed to `common_buf` to make it more visible.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74544/new/

https://reviews.llvm.org/D74544