[PATCH] D29633: [AMDGPU] Distinguish between S/VGPR allocation and encoding granularities

[Konstantin Zhuravlyov via Phabricator via llvm-commits]

kzhuravl created this revision.
Herald added a reviewer: tstellarAMD.
Herald added subscribers: tpr, tony-tye, yaxunl, nhaehnle, wdng.

Tested in https://reviews.llvm.org/D28760.


https://reviews.llvm.org/D29633

Files:
  lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
  lib/Target/AMDGPU/AMDGPUSubtarget.h


Index: lib/Target/AMDGPU/AMDGPUSubtarget.h
===================================================================
--- lib/Target/AMDGPU/AMDGPUSubtarget.h
+++ lib/Target/AMDGPU/AMDGPUSubtarget.h
@@ -644,6 +644,13 @@
 
   /// \returns SGPR allocation granularity supported by the subtarget.
   unsigned getSGPRAllocGranule() const {
+    if (getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS)
+      return 16;
+    return 8;
+  }
+
+  /// \returns SGPR encoding granularity supported by the subtarget.
+  unsigned getSGPREncodingGranule() const {
     return 8;
   }
 
@@ -689,6 +696,11 @@
     return 4;
   }
 
+  /// \returns VGPR encoding granularity supported by the subtarget.
+  unsigned getVGPREncodingGranule() const {
+    return getVGPRAllocGranule();
+  }
+
   /// \returns Total number of VGPRs supported by the subtarget.
   unsigned getTotalNumVGPRs() const {
     return 256;
Index: lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
===================================================================
--- lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
+++ lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
@@ -560,13 +560,13 @@
 
   // SGPRBlocks is actual number of SGPR blocks minus 1.
   ProgInfo.SGPRBlocks = alignTo(ProgInfo.NumSGPRsForWavesPerEU,
-                                STM.getSGPRAllocGranule());
-  ProgInfo.SGPRBlocks = ProgInfo.SGPRBlocks / STM.getSGPRAllocGranule() - 1;
+                                STM.getSGPREncodingGranule());
+  ProgInfo.SGPRBlocks = ProgInfo.SGPRBlocks / STM.getSGPREncodingGranule() - 1;
 
   // VGPRBlocks is actual number of VGPR blocks minus 1.
   ProgInfo.VGPRBlocks = alignTo(ProgInfo.NumVGPRsForWavesPerEU,
-                                STM.getVGPRAllocGranule());
-  ProgInfo.VGPRBlocks = ProgInfo.VGPRBlocks / STM.getVGPRAllocGranule() - 1;
+                                STM.getVGPREncodingGranule());
+  ProgInfo.VGPRBlocks = ProgInfo.VGPRBlocks / STM.getVGPREncodingGranule() - 1;
 
   // Set the value to initialize FP_ROUND and FP_DENORM parts of the mode
   // register.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D29633.87398.patch
Type: text/x-patch
Size: 2023 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170207/ece01782/attachment.bin>