[llvm] r303767 - [AMDGPU] Prevent too large store merges in AMDGPU Subtargets. NFCI.

[Nirav Dave via llvm-commits]

Author: niravd
Date: Wed May 24 10:59:09 2017
New Revision: 303767

URL: http://llvm.org/viewvc/llvm-project?rev=303767&view=rev
Log:
[AMDGPU] Prevent too large store merges in AMDGPU Subtargets. NFCI.

Various address spaces on the SI and R600 subtargets have stricter
limits on memory access size that other address spaces. Use
canMergeStoresTo predicate to prevent the DAGCombiner from creating
these stores as they will be split up during legalization.

Modified:
    llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.cpp
    llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.h
    llvm/trunk/lib/Target/AMDGPU/SIISelLowering.cpp
    llvm/trunk/lib/Target/AMDGPU/SIISelLowering.h

Modified: llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.cpp?rev=303767&r1=303766&r2=303767&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.cpp Wed May 24 10:59:09 2017
@@ -1618,6 +1618,14 @@ EVT R600TargetLowering::getSetCCResultTy
    return VT.changeVectorElementTypeToInteger();
 }
 
+bool R600TargetLowering::canMergeStoresTo(unsigned AS, EVT MemVT) const {
+  // Local and Private addresses do not handle vectors. Limit to i32
+  if ((AS == AMDGPUASI.LOCAL_ADDRESS || AS == AMDGPUASI.PRIVATE_ADDRESS)) {
+    return (MemVT.getSizeInBits() <= 32);
+  }
+  return true;
+}
+
 bool R600TargetLowering::allowsMisalignedMemoryAccesses(EVT VT,
                                                         unsigned AddrSpace,
                                                         unsigned Align,

Modified: llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.h?rev=303767&r1=303766&r2=303767&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.h (original)
+++ llvm/trunk/lib/Target/AMDGPU/R600ISelLowering.h Wed May 24 10:59:09 2017
@@ -44,6 +44,8 @@ public:
   EVT getSetCCResultType(const DataLayout &DL, LLVMContext &,
                          EVT VT) const override;
 
+  bool canMergeStoresTo(unsigned AS, EVT MemVT) const override;
+
   bool allowsMisalignedMemoryAccesses(EVT VT, unsigned AS,
                                       unsigned Align,
                                       bool *IsFast) const override;

Modified: llvm/trunk/lib/Target/AMDGPU/SIISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/SIISelLowering.cpp?rev=303767&r1=303766&r2=303767&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/SIISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/AMDGPU/SIISelLowering.cpp Wed May 24 10:59:09 2017
@@ -698,6 +698,18 @@ bool SITargetLowering::isLegalAddressing
   }
 }
 
+bool SITargetLowering::canMergeStoresTo(unsigned AS, EVT MemVT) const {
+  if (AS == AMDGPUASI.GLOBAL_ADDRESS || AS == AMDGPUASI.FLAT_ADDRESS) {
+    return (MemVT.getSizeInBits() <= 4 * 32);
+  } else if (AS == AMDGPUASI.PRIVATE_ADDRESS) {
+    unsigned MaxPrivateBits = 8 * getSubtarget()->getMaxPrivateElementSize();
+    return (MemVT.getSizeInBits() <= MaxPrivateBits);
+  } else if (AS == AMDGPUASI.LOCAL_ADDRESS) {
+    return (MemVT.getSizeInBits() <= 2 * 32);
+  }
+  return true;
+}
+
 bool SITargetLowering::allowsMisalignedMemoryAccesses(EVT VT,
                                                       unsigned AddrSpace,
                                                       unsigned Align,

Modified: llvm/trunk/lib/Target/AMDGPU/SIISelLowering.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/SIISelLowering.h?rev=303767&r1=303766&r2=303767&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/SIISelLowering.h (original)
+++ llvm/trunk/lib/Target/AMDGPU/SIISelLowering.h Wed May 24 10:59:09 2017
@@ -150,6 +150,8 @@ public:
   bool isLegalAddressingMode(const DataLayout &DL, const AddrMode &AM, Type *Ty,
                              unsigned AS) const override;
 
+  bool canMergeStoresTo(unsigned AS, EVT MemVT) const override;
+
   bool allowsMisalignedMemoryAccesses(EVT VT, unsigned AS,
                                       unsigned Align,
                                       bool *IsFast) const override;