[PATCH] D40578: AMDGPU: Make hazard recognizer aware of maximum clause sizes

Austin Kerbow via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Sat Apr 11 23:57:38 PDT 2020

Previous message: [PATCH] D77965: [mlir][Linalg] NFC: Add utility function to tile, fuse and set marker to use loop.parallel.
Next message: [PATCH] D77967: [Dominators] Facilitate updates to MachineDominator trees
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

kerbowa added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp:60
+  //
+  // LGKMCNT is always 4 bits for an SMEM clause.
+
----------------
gfx10 has 6 bits for LGKMCNT.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D40578/new/

https://reviews.llvm.org/D40578

Previous message: [PATCH] D77965: [mlir][Linalg] NFC: Add utility function to tile, fuse and set marker to use loop.parallel.
Next message: [PATCH] D77967: [Dominators] Facilitate updates to MachineDominator trees
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

More information about the llvm-commits mailing list