summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/AMDGPU/32-bit-local-address-space.ll
diff options
context:
space:
mode:
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/32-bit-local-address-space.ll')
-rw-r--r--llvm/test/CodeGen/AMDGPU/32-bit-local-address-space.ll24
1 files changed, 12 insertions, 12 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/32-bit-local-address-space.ll b/llvm/test/CodeGen/AMDGPU/32-bit-local-address-space.ll
index edad18e244d..ca661cf9a71 100644
--- a/llvm/test/CodeGen/AMDGPU/32-bit-local-address-space.ll
+++ b/llvm/test/CodeGen/AMDGPU/32-bit-local-address-space.ll
@@ -13,7 +13,7 @@
; FUNC-LABEL: {{^}}local_address_load:
; SI: v_mov_b32_e{{32|64}} [[PTR:v[0-9]]]
; SI: ds_read_b32 v{{[0-9]+}}, [[PTR]]
-define void @local_address_load(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
+define amdgpu_kernel void @local_address_load(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
entry:
%0 = load i32, i32 addrspace(3)* %in
store i32 %0, i32 addrspace(1)* %out
@@ -24,7 +24,7 @@ entry:
; SI: s_add_i32 [[SPTR:s[0-9]]]
; SI: v_mov_b32_e32 [[VPTR:v[0-9]+]], [[SPTR]]
; SI: ds_read_b32 [[VPTR]]
-define void @local_address_gep(i32 addrspace(1)* %out, i32 addrspace(3)* %in, i32 %offset) {
+define amdgpu_kernel void @local_address_gep(i32 addrspace(1)* %out, i32 addrspace(3)* %in, i32 %offset) {
entry:
%0 = getelementptr i32, i32 addrspace(3)* %in, i32 %offset
%1 = load i32, i32 addrspace(3)* %0
@@ -35,7 +35,7 @@ entry:
; FUNC-LABEL: {{^}}local_address_gep_const_offset:
; SI: v_mov_b32_e32 [[VPTR:v[0-9]+]], s{{[0-9]+}}
; SI: ds_read_b32 v{{[0-9]+}}, [[VPTR]] offset:4
-define void @local_address_gep_const_offset(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
+define amdgpu_kernel void @local_address_gep_const_offset(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
entry:
%0 = getelementptr i32, i32 addrspace(3)* %in, i32 1
%1 = load i32, i32 addrspace(3)* %0
@@ -48,7 +48,7 @@ entry:
; SI: s_add_i32 [[SPTR:s[0-9]]], s{{[0-9]+}}, 0x10004
; SI: v_mov_b32_e32 [[VPTR:v[0-9]+]], [[SPTR]]
; SI: ds_read_b32 [[VPTR]]
-define void @local_address_gep_large_const_offset(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
+define amdgpu_kernel void @local_address_gep_large_const_offset(i32 addrspace(1)* %out, i32 addrspace(3)* %in) {
entry:
%0 = getelementptr i32, i32 addrspace(3)* %in, i32 16385
%1 = load i32, i32 addrspace(3)* %0
@@ -60,7 +60,7 @@ entry:
; SI: v_cmp_ne_u32
; SI-NOT: v_cmp_ne_u32
; SI: v_cndmask_b32
-define void @null_32bit_lds_ptr(i32 addrspace(1)* %out, i32 addrspace(3)* %lds) nounwind {
+define amdgpu_kernel void @null_32bit_lds_ptr(i32 addrspace(1)* %out, i32 addrspace(3)* %lds) nounwind {
%cmp = icmp ne i32 addrspace(3)* %lds, null
%x = select i1 %cmp, i32 123, i32 456
store i32 %x, i32 addrspace(1)* %out
@@ -71,7 +71,7 @@ define void @null_32bit_lds_ptr(i32 addrspace(1)* %out, i32 addrspace(3)* %lds)
; SI: s_mul_i32
; SI-NEXT: s_add_i32
; SI: ds_read_b32
-define void @mul_32bit_ptr(float addrspace(1)* %out, [3 x float] addrspace(3)* %lds, i32 %tid) {
+define amdgpu_kernel void @mul_32bit_ptr(float addrspace(1)* %out, [3 x float] addrspace(3)* %lds, i32 %tid) {
%ptr = getelementptr [3 x float], [3 x float] addrspace(3)* %lds, i32 %tid, i32 0
%val = load float, float addrspace(3)* %ptr
store float %val, float addrspace(1)* %out
@@ -83,7 +83,7 @@ define void @mul_32bit_ptr(float addrspace(1)* %out, [3 x float] addrspace(3)* %
; FUNC-LABEL: {{^}}infer_ptr_alignment_global_offset:
; SI: v_mov_b32_e32 [[REG:v[0-9]+]], 0
; SI: ds_read_b32 v{{[0-9]+}}, [[REG]]
-define void @infer_ptr_alignment_global_offset(float addrspace(1)* %out, i32 %tid) {
+define amdgpu_kernel void @infer_ptr_alignment_global_offset(float addrspace(1)* %out, i32 %tid) {
%val = load float, float addrspace(3)* @g_lds
store float %val, float addrspace(1)* %out
ret void
@@ -95,14 +95,14 @@ define void @infer_ptr_alignment_global_offset(float addrspace(1)* %out, i32 %ti
; FUNC-LABEL: {{^}}global_ptr:
; SI: ds_write_b32
-define void @global_ptr() nounwind {
+define amdgpu_kernel void @global_ptr() nounwind {
store i32 addrspace(3)* getelementptr ([16383 x i32], [16383 x i32] addrspace(3)* @dst, i32 0, i32 16), i32 addrspace(3)* addrspace(3)* @ptr
ret void
}
; FUNC-LABEL: {{^}}local_address_store:
; SI: ds_write_b32
-define void @local_address_store(i32 addrspace(3)* %out, i32 %val) {
+define amdgpu_kernel void @local_address_store(i32 addrspace(3)* %out, i32 %val) {
store i32 %val, i32 addrspace(3)* %out
ret void
}
@@ -111,7 +111,7 @@ define void @local_address_store(i32 addrspace(3)* %out, i32 %val) {
; SI: s_add_i32 [[SADDR:s[0-9]+]],
; SI: v_mov_b32_e32 [[ADDR:v[0-9]+]], [[SADDR]]
; SI: ds_write_b32 [[ADDR]], v{{[0-9]+}}
-define void @local_address_gep_store(i32 addrspace(3)* %out, i32, i32 %val, i32 %offset) {
+define amdgpu_kernel void @local_address_gep_store(i32 addrspace(3)* %out, i32, i32 %val, i32 %offset) {
%gep = getelementptr i32, i32 addrspace(3)* %out, i32 %offset
store i32 %val, i32 addrspace(3)* %gep, align 4
ret void
@@ -121,7 +121,7 @@ define void @local_address_gep_store(i32 addrspace(3)* %out, i32, i32 %val, i32
; SI: v_mov_b32_e32 [[VPTR:v[0-9]+]], s{{[0-9]+}}
; SI: v_mov_b32_e32 [[VAL:v[0-9]+]], s{{[0-9]+}}
; SI: ds_write_b32 [[VPTR]], [[VAL]] offset:4
-define void @local_address_gep_const_offset_store(i32 addrspace(3)* %out, i32 %val) {
+define amdgpu_kernel void @local_address_gep_const_offset_store(i32 addrspace(3)* %out, i32 %val) {
%gep = getelementptr i32, i32 addrspace(3)* %out, i32 1
store i32 %val, i32 addrspace(3)* %gep, align 4
ret void
@@ -132,7 +132,7 @@ define void @local_address_gep_const_offset_store(i32 addrspace(3)* %out, i32 %v
; SI: s_add_i32 [[SPTR:s[0-9]]], s{{[0-9]+}}, 0x10004
; SI: v_mov_b32_e32 [[VPTR:v[0-9]+]], [[SPTR]]
; SI: ds_write_b32 [[VPTR]], v{{[0-9]+$}}
-define void @local_address_gep_large_const_offset_store(i32 addrspace(3)* %out, i32 %val) {
+define amdgpu_kernel void @local_address_gep_large_const_offset_store(i32 addrspace(3)* %out, i32 %val) {
%gep = getelementptr i32, i32 addrspace(3)* %out, i32 16385
store i32 %val, i32 addrspace(3)* %gep, align 4
ret void
OpenPOWER on IntegriCloud