[PATCH] D133985: [mlir][transform] Add multi-buffering to the transform dialect
Kirsten Lee via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Mon Sep 19 09:41:30 PDT 2022
kile added inline comments.
================
Comment at: mlir/test/Dialect/MemRef/transform-ops.mlir:18
+
+ // CHECK: scf.for %[[IV:.*]] = %[[C0]]
+ scf.for %i0 = %c0 to %c16 step %c4 {
----------------
nicolasvasilache wrote:
> should this capture the fact that the step should has doubled from 4 to 8 ?
Hi, no the loop and its stepping remains the same. All that changes is that `%tmp` is now 2x its original size and all references to `%tmp` now go through a subview of `%tmp`.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D133985/new/
https://reviews.llvm.org/D133985
More information about the llvm-commits
mailing list