[PATCH] D39205: AMDGPU: Initialize WavefrontSize from TD files

[Konstantin Zhuravlyov via Phabricator via llvm-commits]

kzhuravl created this revision.
Herald added subscribers: t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng.

https://reviews.llvm.org/D39205

Files:
  lib/Target/AMDGPU/AMDGPUSubtarget.cpp
  lib/Target/AMDGPU/Processors.td


Index: lib/Target/AMDGPU/Processors.td
===================================================================
--- lib/Target/AMDGPU/Processors.td
+++ lib/Target/AMDGPU/Processors.td
@@ -13,7 +13,7 @@
 // The code produced for "generic" is only useful for tests and cannot
 // reasonably be expected to execute on any particular target.
 def : ProcessorModel<"generic", NoSchedModel, [
-  FeatureGCN
+  FeatureGCN, FeatureWavefrontSize64
 ]>;
 
 //===----------------------------------------------------------------------===//
Index: lib/Target/AMDGPU/AMDGPUSubtarget.cpp
===================================================================
--- lib/Target/AMDGPU/AMDGPUSubtarget.cpp
+++ lib/Target/AMDGPU/AMDGPUSubtarget.cpp
@@ -96,7 +96,7 @@
     TargetTriple(TT),
     Gen(TT.getArch() == Triple::amdgcn ? SOUTHERN_ISLANDS : R600),
     IsaVersion(ISAVersion0_0_0),
-    WavefrontSize(64),
+    WavefrontSize(0),
     LocalMemorySize(0),
     LDSBankCount(0),
     MaxPrivateElementSize(0),


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D39205.119934.patch
Type: text/x-patch
Size: 988 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20171023/db1b7134/attachment.bin>