[PATCH] D23687: AMDGPU/SI: Use a better method for determining the largest pressure sets

Matt Arsenault via llvm-commits llvm-commits at lists.llvm.org
Fri Aug 19 12:58:21 PDT 2016


arsenm accepted this revision.
arsenm added a comment.
This revision is now accepted and ready to land.

LGTM


================
Comment at: lib/Target/AMDGPU/SIRegisterInfo.cpp:111
@@ +110,3 @@
+    for (unsigned j = 0; PSets[j] != -1; ++j) {
+      PressureSetRegUnits[PSets[j]]++;
+    }
----------------
Preincrement

================
Comment at: lib/Target/AMDGPU/SIRegisterInfo.h:190-195
@@ -189,1 +189,8 @@
 
+  bool isSGPRPressureSet(unsigned SetID) const {
+    return SGPRPressureSets.test(SetID) && !VGPRPressureSets.test(SetID);
+  }
+  bool isVGPRPressureSet(unsigned SetID) const {
+    return VGPRPressureSets.test(SetID) && !SGPRPressureSets.test(SetID);
+  }
+
----------------
Why do these need to check that it's not in the other set? Are there still issues from VS_32 being allocatable?


https://reviews.llvm.org/D23687





More information about the llvm-commits mailing list