summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/AMDGPU/local-memory.ll
diff options
context:
space:
mode:
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/local-memory.ll')
-rw-r--r--llvm/test/CodeGen/AMDGPU/local-memory.ll58
1 files changed, 9 insertions, 49 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/local-memory.ll b/llvm/test/CodeGen/AMDGPU/local-memory.ll
index 8d48f594b23..1a11332f865 100644
--- a/llvm/test/CodeGen/AMDGPU/local-memory.ll
+++ b/llvm/test/CodeGen/AMDGPU/local-memory.ll
@@ -1,57 +1,20 @@
-; RUN: llc -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs < %s | FileCheck -check-prefix=CI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
@local_memory.local_mem = internal unnamed_addr addrspace(3) global [128 x i32] undef, align 4
-
-; Check that the LDS size emitted correctly
-; EG: .long 166120
-; EG-NEXT: .long 128
-; SI: .long 47180
-; SI-NEXT: .long 65668
-; CI: .long 47180
-; CI-NEXT: .long 32900
-
-; FUNC-LABEL: {{^}}local_memory:
-
-; EG: LDS_WRITE
-; SI-NOT: s_wqm_b64
-; SI: ds_write_b32
-
-; GROUP_BARRIER must be the last instruction in a clause
-; EG: GROUP_BARRIER
-; EG-NEXT: ALU clause
-; SI: s_barrier
-
-; EG: LDS_READ_RET
-; SI: ds_read_b32 {{v[0-9]+}},
-
-define void @local_memory(i32 addrspace(1)* %out) {
-entry:
- %y.i = call i32 @llvm.r600.read.tidig.x() #0
- %arrayidx = getelementptr inbounds [128 x i32], [128 x i32] addrspace(3)* @local_memory.local_mem, i32 0, i32 %y.i
- store i32 %y.i, i32 addrspace(3)* %arrayidx, align 4
- %add = add nsw i32 %y.i, 1
- %cmp = icmp eq i32 %add, 16
- %.add = select i1 %cmp, i32 0, i32 %add
- call void @llvm.AMDGPU.barrier.local()
- %arrayidx1 = getelementptr inbounds [128 x i32], [128 x i32] addrspace(3)* @local_memory.local_mem, i32 0, i32 %.add
- %0 = load i32, i32 addrspace(3)* %arrayidx1, align 4
- %arrayidx2 = getelementptr inbounds i32, i32 addrspace(1)* %out, i32 %y.i
- store i32 %0, i32 addrspace(1)* %arrayidx2, align 4
- ret void
-}
-
@lds = addrspace(3) global [512 x i32] undef, align 4
-; On SI we need to make sure that the base offset is a register and not
-; an immediate.
+; On SI we need to make sure that the base offset is a register and
+; not an immediate.
+
; FUNC-LABEL: {{^}}load_i32_local_const_ptr:
; GCN: v_mov_b32_e32 v[[ZERO:[0-9]+]], 0
; GCN: ds_read_b32 v{{[0-9]+}}, v[[ZERO]] offset:4
+
; R600: LDS_READ_RET
-define void @load_i32_local_const_ptr(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
+define void @load_i32_local_const_ptr(i32 addrspace(1)* %out, i32 addrspace(3)* %in) #0 {
entry:
%tmp0 = getelementptr [512 x i32], [512 x i32] addrspace(3)* @lds, i32 0, i32 1
%tmp1 = load i32, i32 addrspace(3)* %tmp0
@@ -67,7 +30,7 @@ entry:
; R600: LDS_READ_RET
; GCN-DAG: ds_read_b32
; GCN-DAG: ds_read2_b32
-define void @load_i32_v2i32_local(<2 x i32> addrspace(1)* %out, i32 addrspace(3)* %in) {
+define void @load_i32_v2i32_local(<2 x i32> addrspace(1)* %out, i32 addrspace(3)* %in) #0 {
%scalar = load i32, i32 addrspace(3)* %in
%tmp0 = bitcast i32 addrspace(3)* %in to <2 x i32> addrspace(3)*
%vec_ptr = getelementptr <2 x i32>, <2 x i32> addrspace(3)* %tmp0, i32 2
@@ -78,7 +41,4 @@ define void @load_i32_v2i32_local(<2 x i32> addrspace(1)* %out, i32 addrspace(3)
ret void
}
-declare i32 @llvm.r600.read.tidig.x() #0
-declare void @llvm.AMDGPU.barrier.local()
-
-attributes #0 = { readnone }
+attributes #0 = { nounwind }
OpenPOWER on IntegriCloud