[PATCH] D73634: [AMDGPU] Cluster FLAT instructions with both vaddr and saddr

[Jay Foad via Phabricator via llvm-commits]

This revision was automatically updated to reflect the committed changes.
Closed by commit rGd07a7895792a: [AMDGPU] Cluster FLAT instructions with both vaddr and saddr (authored by foad).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D73634/new/

https://reviews.llvm.org/D73634

Files:
  llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
  llvm/test/CodeGen/AMDGPU/global-saddr.ll
  llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll


Index: llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
===================================================================
--- llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
+++ llvm/test/CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll
@@ -264,11 +264,11 @@
 ; CI-NEXT: buffer_store_dword v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64 offset:52{{$}}
 
 ; GFX9: global_load_dword {{v[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:12
+; GFX9: global_load_dword {{v[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:28
 ; GFX9: global_load_dword {{v[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:44
+
 ; GFX9: global_store_dword v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}}{{$}}
 ; GFX9: global_store_dword v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}} offset:20
-; GFX9: global_load_dword {{v[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:28
-
 ; GFX9: global_store_dword v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}} offset:36
 ; GFX9: global_store_dword v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, s{{\[[0-9]+:[0-9]+\]}} offset:52
 
Index: llvm/test/CodeGen/AMDGPU/global-saddr.ll
===================================================================
--- llvm/test/CodeGen/AMDGPU/global-saddr.ll
+++ llvm/test/CodeGen/AMDGPU/global-saddr.ll
@@ -2,11 +2,11 @@
 
 ; Test for a conv2d like sequence of loads.
 
-; GFX9: global_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:16{{$}}
 ; GFX9: global_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}{{$}}
+; GFX9: global_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:16{{$}}
 ; GFX9: global_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:-16{{$}}
-; GFX9: global_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:32{{$}}
 ; GFX9: global_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:-32{{$}}
+; GFX9: global_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:32{{$}}
 ; GFX9: global_store_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}} offset:8{{$}}
 
 define hidden amdgpu_kernel void @simpleSaddrs(i64 addrspace(1)* %dst_image, i64 addrspace(1)* %src_image ) {
@@ -45,9 +45,9 @@
   store volatile i64 %add7, i64 addrspace(1)* %ptr9
 
 ; Test various offset boundaries.
-; GFX9: global_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}] offset:4088{{$}}
 ; GFX9: global_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, off offset:4088{{$}}
 ; GFX9: global_load_dwordx4 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}] offset:2040{{$}}
+; GFX9: global_load_dwordx2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, s[{{[0-9]+}}:{{[0-9]+}}] offset:4088{{$}}
   %gep11 = getelementptr inbounds i64, i64 addrspace(1)* %gep, i64 511
   %load11 = load i64, i64 addrspace(1)* %gep11
   %gep12 = getelementptr inbounds i64, i64 addrspace(1)* %gep, i64 1023
Index: llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
===================================================================
--- llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
+++ llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
@@ -365,23 +365,14 @@
   }
 
   if (isFLAT(LdSt)) {
-    const MachineOperand *VAddr = getNamedOperand(LdSt, AMDGPU::OpName::vaddr);
-    if (VAddr) {
-      // Can't analyze 2 offsets.
-      // FIXME remove this restriction!
-      if (getNamedOperand(LdSt, AMDGPU::OpName::saddr))
-        return false;
-
-      BaseOp = VAddr;
-    } else {
-      // scratch instructions have either vaddr or saddr.
-      BaseOp = getNamedOperand(LdSt, AMDGPU::OpName::saddr);
-    }
-
+    // Instructions have either vaddr or saddr or both.
+    BaseOp = getNamedOperand(LdSt, AMDGPU::OpName::vaddr);
+    if (BaseOp)
+      BaseOps.push_back(BaseOp);
+    BaseOp = getNamedOperand(LdSt, AMDGPU::OpName::saddr);
+    if (BaseOp)
+      BaseOps.push_back(BaseOp);
     Offset = getNamedOperand(LdSt, AMDGPU::OpName::offset)->getImm();
-    if (!BaseOp->isReg())
-      return false;
-    BaseOps.push_back(BaseOp);
     return true;
   }
 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D73634.241179.patch
Type: text/x-patch
Size: 4357 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200129/4605b9d5/attachment.bin>