[PATCH] D121437: [AMDGPU] Add s_nop WaitStates between neighboring mfma

Stanislav Mekhanoshin via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Fri Mar 11 10:19:33 PST 2022


rampitec added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp:270
+                  TSchedModel.getWriteProcResEnd(SC))) {
+    if (!strcmp(TSchedModel.getProcResource(PRE.ProcResourceIdx)->Name,
+                "HWXDL")) {
----------------
If you must compare string it is better to find this MCProcResourceDesc once and then compare the pointer.


================
Comment at: llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp:44
+    cl::desc(
+        "Percent of wait states between neighboring MFMA to pad with s_nops"));
+
----------------
arsenm wrote:
> I'm not sure what a percentage means here
Yep, it shall have a text this is percentage of the mfma latency.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D121437/new/

https://reviews.llvm.org/D121437



More information about the llvm-commits mailing list