diff options
author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2016-04-14 22:11:51 +0000 |
---|---|---|
committer | Matt Arsenault <Matthew.Arsenault@amd.com> | 2016-04-14 22:11:51 +0000 |
commit | fd8ab09c0e5587621ce4d7dc5c0307130fa4b589 (patch) | |
tree | f1129a49b2935996f402a0702702167be11501fe | |
parent | 16f13e252b2b41979abb563aa0e9a59cbb60e88e (diff) | |
download | bcm5719-llvm-fd8ab09c0e5587621ce4d7dc5c0307130fa4b589.tar.gz bcm5719-llvm-fd8ab09c0e5587621ce4d7dc5c0307130fa4b589.zip |
AMDGPU: Include LDS size in printed comment
llvm-svn: 266382
-rw-r--r-- | llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp | 2 | ||||
-rw-r--r-- | llvm/test/CodeGen/AMDGPU/lds-size.ll | 14 |
2 files changed, 12 insertions, 4 deletions
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp index 2673a331fd8..299872d3dee 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp @@ -232,6 +232,8 @@ bool AMDGPUAsmPrinter::runOnMachineFunction(MachineFunction &MF) { false); OutStreamer->emitRawComment(" ScratchSize: " + Twine(KernelInfo.ScratchSize), false); + OutStreamer->emitRawComment(" LDSByteSize: " + Twine(KernelInfo.LDSSize) + + " bytes/workgroup (compile time only)", false); OutStreamer->emitRawComment(" COMPUTE_PGM_RSRC2:USER_SGPR: " + Twine(G_00B84C_USER_SGPR(KernelInfo.ComputePGMRSrc2)), diff --git a/llvm/test/CodeGen/AMDGPU/lds-size.ll b/llvm/test/CodeGen/AMDGPU/lds-size.ll index 3e8328659fd..1607713090e 100644 --- a/llvm/test/CodeGen/AMDGPU/lds-size.ll +++ b/llvm/test/CodeGen/AMDGPU/lds-size.ll @@ -1,11 +1,17 @@ -; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s +; RUN: llc -march=amdgcn < %s | FileCheck -check-prefix=ALL -check-prefix=GCN %s +; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=ALL -check-prefix=EG %s ; This test makes sure we do not double count global values when they are ; used in different basic blocks. -; CHECK: .long 166120 -; CHECK-NEXT: .long 1 -; CHECK-LABEL: {{^}}test: +; GCN: .long 47180 +; GCN-NEXT: .long 32900 + +; EG: .long 166120 +; EG-NEXT: .long 1 +; ALL: {{^}}test: + +; GCN: ; LDSByteSize: 4 bytes/workgroup (compile time only) @lds = internal unnamed_addr addrspace(3) global i32 undef, align 4 define void @test(i32 addrspace(1)* %out, i32 %cond) { |