aboutsummaryrefslogtreecommitdiffstats
path: root/test/CodeGen/R600/local-memory-two-objects.ll
diff options
context:
space:
mode:
authorPirama Arumuga Nainar <pirama@google.com>2015-04-10 22:08:18 +0000
committerAndroid Git Automerger <android-git-automerger@android.com>2015-04-10 22:08:18 +0000
commit13a7db5b9c4f5e543d037be68ec3428216bfd550 (patch)
tree1b2c9792582e12f5af0b1512e3094425f0dc0df9 /test/CodeGen/R600/local-memory-two-objects.ll
parent0eb46f5d1e06a4284663d636a74b06adc3a161d7 (diff)
parent31195f0bdca6ee2a5e72d07edf13e1d81206d949 (diff)
downloadexternal_llvm-13a7db5b9c4f5e543d037be68ec3428216bfd550.zip
external_llvm-13a7db5b9c4f5e543d037be68ec3428216bfd550.tar.gz
external_llvm-13a7db5b9c4f5e543d037be68ec3428216bfd550.tar.bz2
am 31195f0b: Merge "Update aosp/master llvm for rebase to r233350"
* commit '31195f0bdca6ee2a5e72d07edf13e1d81206d949': Update aosp/master llvm for rebase to r233350
Diffstat (limited to 'test/CodeGen/R600/local-memory-two-objects.ll')
-rw-r--r--test/CodeGen/R600/local-memory-two-objects.ll19
1 files changed, 10 insertions, 9 deletions
diff --git a/test/CodeGen/R600/local-memory-two-objects.ll b/test/CodeGen/R600/local-memory-two-objects.ll
index 3d90ab1..caa4b19 100644
--- a/test/CodeGen/R600/local-memory-two-objects.ll
+++ b/test/CodeGen/R600/local-memory-two-objects.ll
@@ -5,7 +5,6 @@
@local_memory_two_objects.local_mem0 = internal unnamed_addr addrspace(3) global [4 x i32] undef, align 4
@local_memory_two_objects.local_mem1 = internal unnamed_addr addrspace(3) global [4 x i32] undef, align 4
-; EG: {{^}}local_memory_two_objects:
; Check that the LDS size emitted correctly
; EG: .long 166120
@@ -13,6 +12,8 @@
; GCN: .long 47180
; GCN-NEXT: .long 38792
+; EG: {{^}}local_memory_two_objects:
+
; We would like to check the the lds writes are using different
; addresses, but due to variations in the scheduler, we can't do
; this consistently on evergreen GPUs.
@@ -37,21 +38,21 @@
define void @local_memory_two_objects(i32 addrspace(1)* %out) {
entry:
%x.i = call i32 @llvm.r600.read.tidig.x() #0
- %arrayidx = getelementptr inbounds [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem0, i32 0, i32 %x.i
+ %arrayidx = getelementptr inbounds [4 x i32], [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem0, i32 0, i32 %x.i
store i32 %x.i, i32 addrspace(3)* %arrayidx, align 4
%mul = shl nsw i32 %x.i, 1
- %arrayidx1 = getelementptr inbounds [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem1, i32 0, i32 %x.i
+ %arrayidx1 = getelementptr inbounds [4 x i32], [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem1, i32 0, i32 %x.i
store i32 %mul, i32 addrspace(3)* %arrayidx1, align 4
%sub = sub nsw i32 3, %x.i
call void @llvm.AMDGPU.barrier.local()
- %arrayidx2 = getelementptr inbounds [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem0, i32 0, i32 %sub
- %0 = load i32 addrspace(3)* %arrayidx2, align 4
- %arrayidx3 = getelementptr inbounds i32 addrspace(1)* %out, i32 %x.i
+ %arrayidx2 = getelementptr inbounds [4 x i32], [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem0, i32 0, i32 %sub
+ %0 = load i32, i32 addrspace(3)* %arrayidx2, align 4
+ %arrayidx3 = getelementptr inbounds i32, i32 addrspace(1)* %out, i32 %x.i
store i32 %0, i32 addrspace(1)* %arrayidx3, align 4
- %arrayidx4 = getelementptr inbounds [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem1, i32 0, i32 %sub
- %1 = load i32 addrspace(3)* %arrayidx4, align 4
+ %arrayidx4 = getelementptr inbounds [4 x i32], [4 x i32] addrspace(3)* @local_memory_two_objects.local_mem1, i32 0, i32 %sub
+ %1 = load i32, i32 addrspace(3)* %arrayidx4, align 4
%add = add nsw i32 %x.i, 4
- %arrayidx5 = getelementptr inbounds i32 addrspace(1)* %out, i32 %add
+ %arrayidx5 = getelementptr inbounds i32, i32 addrspace(1)* %out, i32 %add
store i32 %1, i32 addrspace(1)* %arrayidx5, align 4
ret void
}