summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll
diff options
context:
space:
mode:
authorMatt Arsenault <Matthew.Arsenault@amd.com>2017-03-21 21:39:51 +0000
committerMatt Arsenault <Matthew.Arsenault@amd.com>2017-03-21 21:39:51 +0000
commit3dbeefa978fb7e7b231b249f9cd90c67b9e83277 (patch)
treed74bf7fe30e44588d573919f3625edacb2586112 /llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll
parentf6021ecddc73d14c94ad70938250d58f330795be (diff)
downloadbcm5719-llvm-3dbeefa978fb7e7b231b249f9cd90c67b9e83277.tar.gz
bcm5719-llvm-3dbeefa978fb7e7b231b249f9cd90c67b9e83277.zip
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel
Currently the default C calling convention functions are treated the same as compute kernels. Make this explicit so the default calling convention can be changed to a non-kernel. Converted with perl -pi -e 's/define void/define amdgpu_kernel void/' on the relevant test directories (and undoing in one place that actually wanted a non-kernel). llvm-svn: 298444
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll')
-rw-r--r--llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll10
1 files changed, 5 insertions, 5 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll b/llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll
index 281e49f804c..601aca48e1e 100644
--- a/llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll
+++ b/llvm/test/CodeGen/AMDGPU/reduce-store-width-alignment.ll
@@ -3,7 +3,7 @@
; GCN-LABEL: {{^}}store_v2i32_as_v4i16_align_4:
; GCN: s_load_dwordx2
; GCN: ds_write2_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset1:1{{$}}
-define void @store_v2i32_as_v4i16_align_4(<4 x i16> addrspace(3)* align 4 %out, <2 x i32> %x) #0 {
+define amdgpu_kernel void @store_v2i32_as_v4i16_align_4(<4 x i16> addrspace(3)* align 4 %out, <2 x i32> %x) #0 {
%x.bc = bitcast <2 x i32> %x to <4 x i16>
store <4 x i16> %x.bc, <4 x i16> addrspace(3)* %out, align 4
ret void
@@ -13,7 +13,7 @@ define void @store_v2i32_as_v4i16_align_4(<4 x i16> addrspace(3)* align 4 %out,
; GCN: s_load_dwordx4
; GCN-DAG: ds_write2_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset0:2 offset1:3
; GCN-DAG: ds_write2_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset1:1{{$}}
-define void @store_v4i32_as_v8i16_align_4(<8 x i16> addrspace(3)* align 4 %out, <4 x i32> %x) #0 {
+define amdgpu_kernel void @store_v4i32_as_v8i16_align_4(<8 x i16> addrspace(3)* align 4 %out, <4 x i32> %x) #0 {
%x.bc = bitcast <4 x i32> %x to <8 x i16>
store <8 x i16> %x.bc, <8 x i16> addrspace(3)* %out, align 4
ret void
@@ -22,7 +22,7 @@ define void @store_v4i32_as_v8i16_align_4(<8 x i16> addrspace(3)* align 4 %out,
; GCN-LABEL: {{^}}store_v2i32_as_i64_align_4:
; GCN: s_load_dwordx2
; GCN: ds_write2_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset1:1{{$}}
-define void @store_v2i32_as_i64_align_4(<4 x i16> addrspace(3)* align 4 %out, <2 x i32> %x) #0 {
+define amdgpu_kernel void @store_v2i32_as_i64_align_4(<4 x i16> addrspace(3)* align 4 %out, <2 x i32> %x) #0 {
%x.bc = bitcast <2 x i32> %x to <4 x i16>
store <4 x i16> %x.bc, <4 x i16> addrspace(3)* %out, align 4
ret void
@@ -32,7 +32,7 @@ define void @store_v2i32_as_i64_align_4(<4 x i16> addrspace(3)* align 4 %out, <2
; GCN: s_load_dwordx4
; GCN-DAG: ds_write2_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset0:2 offset1:3
; GCN-DAG: ds_write2_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset1:1{{$}}
-define void @store_v4i32_as_v2i64_align_4(<2 x i64> addrspace(3)* align 4 %out, <4 x i32> %x) #0 {
+define amdgpu_kernel void @store_v4i32_as_v2i64_align_4(<2 x i64> addrspace(3)* align 4 %out, <4 x i32> %x) #0 {
%x.bc = bitcast <4 x i32> %x to <2 x i64>
store <2 x i64> %x.bc, <2 x i64> addrspace(3)* %out, align 4
ret void
@@ -44,7 +44,7 @@ define void @store_v4i32_as_v2i64_align_4(<2 x i64> addrspace(3)* align 4 %out,
; GCN: buffer_load_ushort
; GCN: buffer_load_ushort
; GCN: ds_write2_b32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} offset1:1{{$}}
-define void @store_v4i16_as_v2i32_align_4(<2 x i32> addrspace(3)* align 4 %out, <4 x i16> %x) #0 {
+define amdgpu_kernel void @store_v4i16_as_v2i32_align_4(<2 x i32> addrspace(3)* align 4 %out, <4 x i16> %x) #0 {
%x.bc = bitcast <4 x i16> %x to <2 x i32>
store <2 x i32> %x.bc, <2 x i32> addrspace(3)* %out, align 4
ret void
OpenPOWER on IntegriCloud