[PATCH] D140174: [OpenMP] Increase max teams in OMPGridValues

[Kevin Sala Penadés via Phabricator via llvm-commits]

kevinsala created this revision.
kevinsala added reviewers: jdoerfert, jhuber6, tianshilei1992, JonChesterfield.
kevinsala added a project: OpenMP.
Herald added subscribers: kosarev, guansong, tpr, yaxunl.
Herald added a project: All.
kevinsala requested review of this revision.
Herald added subscribers: llvm-commits, sstefan1.
Herald added a project: LLVM.

Increasing the maximum number of teams in OMPGridValues to 64K for both NVIDIA and AMDGPU.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D140174

Files:
  llvm/include/llvm/Frontend/OpenMP/OMPGridValues.h


Index: llvm/include/llvm/Frontend/OpenMP/OMPGridValues.h
===================================================================
--- llvm/include/llvm/Frontend/OpenMP/OMPGridValues.h
+++ llvm/include/llvm/Frontend/OpenMP/OMPGridValues.h
@@ -84,7 +84,7 @@
 static constexpr GV AMDGPUGridValues64 = {
     256,  // GV_Slot_Size
     64,   // GV_Warp_Size
-    128,  // GV_Max_Teams
+    (1 << 16),  // GV_Max_Teams
     896,  // GV_SimpleBufferSize
     1024, // GV_Max_WG_Size,
     256,  // GV_Default_WG_Size
@@ -93,7 +93,7 @@
 static constexpr GV AMDGPUGridValues32 = {
     256,  // GV_Slot_Size
     32,   // GV_Warp_Size
-    128,  // GV_Max_Teams
+    (1 << 16),  // GV_Max_Teams
     896,  // GV_SimpleBufferSize
     1024, // GV_Max_WG_Size,
     256,  // GV_Default_WG_Size
@@ -108,7 +108,7 @@
 static constexpr GV NVPTXGridValues = {
     256,  // GV_Slot_Size
     32,   // GV_Warp_Size
-    1024, // GV_Max_Teams
+    (1 << 16), // GV_Max_Teams
     896,  // GV_SimpleBufferSize
     1024, // GV_Max_WG_Size
     128,  // GV_Default_WG_Size


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D140174.483375.patch
Type: text/x-patch
Size: 1049 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20221215/af2aa88e/attachment.bin>