[llvm] r302277 - AMDGPU/AMDHSA: Set COMPUTE_PGM_RSRC2:LDS_SIZE to 0
Konstantin Zhuravlyov via llvm-commits
llvm-commits at lists.llvm.org
Fri May 5 13:13:56 PDT 2017
Author: kzhuravl
Date: Fri May 5 15:13:55 2017
New Revision: 302277
URL: http://llvm.org/viewvc/llvm-project?rev=302277&view=rev
Log:
AMDGPU/AMDHSA: Set COMPUTE_PGM_RSRC2:LDS_SIZE to 0
This field is populated by the CP
Differential Revision: https://reviews.llvm.org/D32619
Modified:
llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
llvm/trunk/test/CodeGen/AMDGPU/lds-size.ll
Modified: llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp?rev=302277&r1=302276&r2=302277&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp (original)
+++ llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp Fri May 5 15:13:55 2017
@@ -713,7 +713,8 @@ void AMDGPUAsmPrinter::getSIProgramInfo(
S_00B84C_TG_SIZE_EN(MFI->hasWorkGroupInfo()) |
S_00B84C_TIDIG_COMP_CNT(TIDIGCompCnt) |
S_00B84C_EXCP_EN_MSB(0) |
- S_00B84C_LDS_SIZE(ProgInfo.LDSBlocks) |
+ // For AMDHSA, LDS_SIZE must be zero, as it is populated by the CP.
+ S_00B84C_LDS_SIZE(STM.isAmdHsaOS() ? 0 : ProgInfo.LDSBlocks) |
S_00B84C_EXCP_EN(0);
}
Modified: llvm/trunk/test/CodeGen/AMDGPU/lds-size.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AMDGPU/lds-size.ll?rev=302277&r1=302276&r2=302277&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/AMDGPU/lds-size.ll (original)
+++ llvm/trunk/test/CodeGen/AMDGPU/lds-size.ll Fri May 5 15:13:55 2017
@@ -1,4 +1,5 @@
; RUN: llc -march=amdgcn < %s | FileCheck -check-prefix=ALL -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa < %s | FileCheck -check-prefix=ALL -check-prefix=HSA %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
@@ -11,6 +12,9 @@
; EG-NEXT: .long 1
; ALL: {{^}}test:
+; HSA: granulated_lds_size = 0
+; HSA: workgroup_group_segment_byte_size = 4
+
; GCN: ; LDSByteSize: 4 bytes/workgroup (compile time only)
@lds = internal unnamed_addr addrspace(3) global i32 undef, align 4
More information about the llvm-commits
mailing list