[AMDGPU] Cleanup in memory legalizer tests. NFC.

llvm-svn: 325042

5 files changed, 531 insertions, 527 deletions

diff --git a/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-cmpxchg.ll b/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-cmpxchg.ll
index 337dbd1e64f..73354053799 100644
--- a/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-cmpxchg.ll
+++ b/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-cmpxchg.ll
@@ -1,11 +1,11 @@
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck %s
-
-; CHECK-LABEL: {{^}}system_monotonic_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX8 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX8 %s
+
+; GCN-LABEL: {{^}}system_monotonic_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @system_monotonic_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -14,11 +14,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_acquire_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_acquire_monotonic:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_acquire_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -27,11 +27,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_release_monotonic
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_release_monotonic:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:    buffer_wbinvl1_vol
 define amdgpu_kernel void @system_release_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -40,11 +40,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_acq_rel_monotonic
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_acq_rel_monotonic:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_acq_rel_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -53,11 +53,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_seq_cst_monotonic
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_seq_cst_monotonic:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_seq_cst_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -66,11 +66,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_acquire_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_acquire_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_acquire_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -79,11 +79,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_release_acquire
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_release_acquire:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_release_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -92,11 +92,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_acq_rel_acquire
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_acq_rel_acquire:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_acq_rel_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -105,11 +105,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_seq_cst_acquire
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_seq_cst_acquire:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_seq_cst_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -118,11 +118,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_seq_cst_seq_cst
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_seq_cst_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_seq_cst_seq_cst(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -131,11 +131,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_monotonic_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_monotonic_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_monotonic_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -144,11 +144,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_acquire_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_acquire_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_acquire_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -157,11 +157,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_release_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_release_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_release_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -170,11 +170,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_acq_rel_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_acq_rel_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_acq_rel_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -183,11 +183,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_seq_cst_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_seq_cst_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_seq_cst_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -196,11 +196,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_acquire_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_acquire_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_acquire_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -209,11 +209,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_release_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_release_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_release_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -222,11 +222,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_acq_rel_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_acq_rel_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_acq_rel_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -235,11 +235,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_seq_cst_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_seq_cst_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_seq_cst_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -248,11 +248,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_seq_cst_seq_cst
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_seq_cst_seq_cst:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_seq_cst_seq_cst(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -261,11 +261,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_monotonic_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_monotonic_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_monotonic_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -274,11 +274,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_acquire_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_acquire_monotonic:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_acquire_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -287,11 +287,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_release_monotonic
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_release_monotonic:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:    buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_release_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -300,11 +300,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_acq_rel_monotonic
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_acq_rel_monotonic:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_acq_rel_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -313,11 +313,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_seq_cst_monotonic
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_seq_cst_monotonic:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_seq_cst_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -326,11 +326,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_acquire_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_acquire_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_acquire_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -339,11 +339,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_release_acquire
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_release_acquire:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_release_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -352,11 +352,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_acq_rel_acquire
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_acq_rel_acquire:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_acq_rel_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -365,11 +365,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_seq_cst_acquire
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_seq_cst_acquire:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_seq_cst_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -378,11 +378,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_seq_cst_seq_cst
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_seq_cst_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_seq_cst_seq_cst(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -391,11 +391,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_monotonic_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_monotonic_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_monotonic_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -404,11 +404,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_acquire_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_acquire_monotonic:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_acquire_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -417,11 +417,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_release_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_release_monotonic:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:    buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_release_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -430,11 +430,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_acq_rel_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_acq_rel_monotonic:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_acq_rel_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -443,11 +443,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_seq_cst_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_seq_cst_monotonic:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_seq_cst_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -456,11 +456,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_acquire_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_acquire_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_acquire_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -469,11 +469,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_release_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_release_acquire:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_release_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -482,11 +482,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_acq_rel_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_acq_rel_acquire:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_acq_rel_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -495,11 +495,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_seq_cst_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_seq_cst_acquire:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_seq_cst_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -508,11 +508,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_seq_cst_seq_cst
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_seq_cst_seq_cst:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_seq_cst_seq_cst(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -521,11 +521,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_monotonic_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_monotonic_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_monotonic_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -534,11 +534,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_acquire_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_acquire_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_acquire_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -547,11 +547,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_release_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_release_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_release_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -560,11 +560,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_acq_rel_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_acq_rel_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_acq_rel_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -573,11 +573,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_seq_cst_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_seq_cst_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_seq_cst_monotonic(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -586,11 +586,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_acquire_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_acquire_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_acquire_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -599,11 +599,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_release_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_release_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_release_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -612,11 +612,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_acq_rel_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_acq_rel_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_acq_rel_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -625,11 +625,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_seq_cst_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_seq_cst_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_seq_cst_acquire(
     i32* %out, i32 %in, i32 %old) {
 entry:
@@ -638,11 +638,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_seq_cst_seq_cst
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_seq_cst_seq_cst:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_cmpswap v[{{[0-9]+\:[0-9]+}}], v[{{[0-9]+}}:{{[0-9]+}}]{{( offset:[0-9]+)*}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_seq_cst_seq_cst(
     i32* %out, i32 %in, i32 %old) {
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-fence.ll b/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-fence.ll
index c8e920a1854..02104826e6a 100644
--- a/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-fence.ll
+++ b/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-fence.ll
@@ -1,8 +1,8 @@
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx600 -verify-machineinstrs < %s | FileCheck -check-prefix=FUNC -check-prefix=GCN -check-prefix=GFX6 %s
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefix=FUNC -check-prefix=GCN -check-prefix=GFX8 %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefix=FUNC -check-prefix=GCN -check-prefix=GFX8 %s
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx600 -verify-machineinstrs < %s | FileCheck -check-prefixes=FUNC,GCN,GFX6,GFX68 %s
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefixes=FUNC,GCN,GFX8,GFX68 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefixes=FUNC,GCN,GFX8,GFX68 %s
 
-; FUNC-LABEL: {{^}}system_acquire
+; FUNC-LABEL: {{^}}system_acquire:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GFX6:       s_waitcnt vmcnt(0){{$}}
@@ -16,7 +16,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}system_release
+; FUNC-LABEL: {{^}}system_release:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_waitcnt vmcnt(0){{$}}
@@ -27,7 +27,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}system_acq_rel
+; FUNC-LABEL: {{^}}system_acq_rel:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_waitcnt vmcnt(0){{$}}
@@ -40,7 +40,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}system_seq_cst
+; FUNC-LABEL: {{^}}system_seq_cst:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_waitcnt vmcnt(0){{$}}
@@ -53,7 +53,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}singlethread_acquire
+; FUNC-LABEL: {{^}}singlethread_acquire:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
@@ -63,7 +63,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}singlethread_release
+; FUNC-LABEL: {{^}}singlethread_release:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
@@ -73,7 +73,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}singlethread_acq_rel
+; FUNC-LABEL: {{^}}singlethread_acq_rel:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
@@ -83,7 +83,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}singlethread_seq_cst
+; FUNC-LABEL: {{^}}singlethread_seq_cst:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
@@ -93,7 +93,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}agent_acquire
+; FUNC-LABEL: {{^}}agent_acquire:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GFX6:       s_waitcnt vmcnt(0){{$}}
@@ -107,7 +107,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}agent_release
+; FUNC-LABEL: {{^}}agent_release:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_waitcnt vmcnt(0){{$}}
@@ -118,7 +118,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}agent_acq_rel
+; FUNC-LABEL: {{^}}agent_acq_rel:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_waitcnt vmcnt(0){{$}}
@@ -131,7 +131,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}agent_seq_cst
+; FUNC-LABEL: {{^}}agent_seq_cst:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_waitcnt vmcnt(0){{$}}
@@ -144,8 +144,9 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}workgroup_acquire
+; FUNC-LABEL: {{^}}workgroup_acquire:
 ; GCN:        %bb.0
+; GFX68-NOT:  s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
 define amdgpu_kernel void @workgroup_acquire() {
@@ -154,8 +155,9 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}workgroup_release
+; FUNC-LABEL: {{^}}workgroup_release:
 ; GCN:        %bb.0
+; GFX68-NOT:  s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
 define amdgpu_kernel void @workgroup_release() {
@@ -164,8 +166,9 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}workgroup_acq_rel
+; FUNC-LABEL: {{^}}workgroup_acq_rel:
 ; GCN:        %bb.0
+; GFX68-NOT:  s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
 define amdgpu_kernel void @workgroup_acq_rel() {
@@ -174,8 +177,9 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}workgroup_seq_cst
+; FUNC-LABEL: {{^}}workgroup_seq_cst:
 ; GCN:        %bb.0
+; GFX68-NOT:  s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
 define amdgpu_kernel void @workgroup_seq_cst() {
@@ -184,7 +188,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}wavefront_acquire
+; FUNC-LABEL: {{^}}wavefront_acquire:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
@@ -194,7 +198,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}wavefront_release
+; FUNC-LABEL: {{^}}wavefront_release:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
@@ -204,7 +208,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}wavefront_acq_rel
+; FUNC-LABEL: {{^}}wavefront_acq_rel:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
@@ -214,7 +218,7 @@ entry:
   ret void
 }
 
-; FUNC-LABEL: {{^}}wavefront_seq_cst
+; FUNC-LABEL: {{^}}wavefront_seq_cst:
 ; GCN:        %bb.0
 ; GCN-NOT:    ATOMIC_FENCE
 ; GCN:        s_endpgm
diff --git a/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-rmw.ll b/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-rmw.ll
index ae265f92cf0..7bf4b93ec84 100644
--- a/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-rmw.ll
+++ b/llvm/test/CodeGen/AMDGPU/memory-legalizer-atomic-rmw.ll
@@ -1,11 +1,11 @@
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX8 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX8 %s
 
-; CHECK-LABEL: {{^}}system_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @system_monotonic(
     i32* %out, i32 %in) {
 entry:
@@ -13,11 +13,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_acquire(
     i32* %out, i32 %in) {
 entry:
@@ -25,11 +25,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_release
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_release:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:    buffer_wbinvl1_vol
 define amdgpu_kernel void @system_release(
     i32* %out, i32 %in) {
 entry:
@@ -37,11 +37,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_acq_rel
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_acq_rel:
+; GCN:         s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:    flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NEXT:    s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @system_acq_rel(
     i32* %out, i32 %in) {
 entry:
@@ -49,11 +49,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}system_seq_cst
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}system_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @system_seq_cst(
     i32* %out, i32 %in) {
 entry:
@@ -61,11 +61,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_monotonic(
     i32* %out, i32 %in) {
 entry:
@@ -73,11 +73,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_acquire(
     i32* %out, i32 %in) {
 entry:
@@ -85,11 +85,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_release
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_release:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_release(
     i32* %out, i32 %in) {
 entry:
@@ -97,11 +97,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_acq_rel
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_acq_rel:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_acq_rel(
     i32* %out, i32 %in) {
 entry:
@@ -109,11 +109,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}singlethread_seq_cst
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}singlethread_seq_cst:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @singlethread_seq_cst(
     i32* %out, i32 %in) {
 entry:
@@ -121,11 +121,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_monotonic(
     i32* %out, i32 %in) {
 entry:
@@ -133,11 +133,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_acquire(
     i32* %out, i32 %in) {
 entry:
@@ -145,11 +145,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_release
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_release:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:    buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_release(
     i32* %out, i32 %in) {
 entry:
@@ -157,11 +157,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_acq_rel
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_acq_rel:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_acq_rel(
     i32* %out, i32 %in) {
 entry:
@@ -169,11 +169,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}agent_seq_cst
-; CHECK:       s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NEXT:  s_waitcnt vmcnt(0){{$}}
-; CHECK-NEXT:  buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}agent_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NEXT:  buffer_wbinvl1_vol
 define amdgpu_kernel void @agent_seq_cst(
     i32* %out, i32 %in) {
 entry:
@@ -181,11 +181,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_monotonic(
     i32* %out, i32 %in) {
 entry:
@@ -193,11 +193,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_acquire(
     i32* %out, i32 %in) {
 entry:
@@ -205,11 +205,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_release
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_release:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:    buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_release(
     i32* %out, i32 %in) {
 entry:
@@ -217,11 +217,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_acq_rel
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_acq_rel:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_acq_rel(
     i32* %out, i32 %in) {
 entry:
@@ -229,11 +229,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}workgroup_seq_cst
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}workgroup_seq_cst:
+; GFX8-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GFX8-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @workgroup_seq_cst(
     i32* %out, i32 %in) {
 entry:
@@ -241,11 +241,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_monotonic
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_monotonic:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_monotonic(
     i32* %out, i32 %in) {
 entry:
@@ -253,11 +253,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_acquire
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_acquire:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_acquire(
     i32* %out, i32 %in) {
 entry:
@@ -265,11 +265,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_release
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_release:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_release(
     i32* %out, i32 %in) {
 entry:
@@ -277,11 +277,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_acq_rel
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_acq_rel:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_acq_rel(
     i32* %out, i32 %in) {
 entry:
@@ -289,11 +289,11 @@ entry:
   ret void
 }
 
-; CHECK-LABEL: {{^}}wavefront_seq_cst
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
-; CHECK-NOT:   s_waitcnt vmcnt(0){{$}}
-; CHECK-NOT:   buffer_wbinvl1_vol
+; GCN-LABEL: {{^}}wavefront_seq_cst:
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN:       flat_atomic_swap v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}{{$}}
+; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
+; GCN-NOT:   buffer_wbinvl1_vol
 define amdgpu_kernel void @wavefront_seq_cst(
     i32* %out, i32 %in) {
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/memory-legalizer-load.ll b/llvm/test/CodeGen/AMDGPU/memory-legalizer-load.ll
index 1b0f00a924f..d64e240d791 100644
--- a/llvm/test/CodeGen/AMDGPU/memory-legalizer-load.ll
+++ b/llvm/test/CodeGen/AMDGPU/memory-legalizer-load.ll
@@ -1,11 +1,11 @@
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX8 %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX8 %s
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX8,GFX89 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX8,GFX89 %s
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX9,GFX89 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX9,GFX89 %s
 
 declare i32 @llvm.amdgcn.workitem.id.x()
 
-; GCN-LABEL: {{^}}system_unordered
+; GCN-LABEL: {{^}}system_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -19,9 +19,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}system_monotonic
+; GCN-LABEL: {{^}}system_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
+; GFX89:     flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:   buffer_wbinvl1_vol
 ; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
@@ -33,12 +33,12 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}system_acquire
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
-; GCN-NEXT:  s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  buffer_wbinvl1_vol
-; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
+; GCN-LABEL: {{^}}system_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX89-NEXT: buffer_wbinvl1_vol
+; GCN:        flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
 define amdgpu_kernel void @system_acquire(
     i32* %in, i32* %out) {
 entry:
@@ -47,12 +47,12 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}system_seq_cst
-; GCN:       s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
-; GCN-NEXT:  s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  buffer_wbinvl1_vol
-; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
+; GCN-LABEL: {{^}}system_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX89-NEXT: buffer_wbinvl1_vol
+; GCN:        flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
 define amdgpu_kernel void @system_seq_cst(
     i32* %in, i32* %out) {
 entry:
@@ -61,7 +61,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_unordered
+; GCN-LABEL: {{^}}singlethread_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -75,7 +75,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_monotonic
+; GCN-LABEL: {{^}}singlethread_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -89,7 +89,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_acquire
+; GCN-LABEL: {{^}}singlethread_acquire:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -103,7 +103,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_seq_cst
+; GCN-LABEL: {{^}}singlethread_seq_cst:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -117,7 +117,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_unordered
+; GCN-LABEL: {{^}}agent_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -131,9 +131,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_monotonic
+; GCN-LABEL: {{^}}agent_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
+; GFX89:     flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:   buffer_wbinvl1_vol
 ; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
@@ -145,12 +145,12 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_acquire
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
-; GCN-NEXT:  s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  buffer_wbinvl1_vol
-; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
+; GCN-LABEL: {{^}}agent_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX89-NEXT: buffer_wbinvl1_vol
+; GCN:        flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
 define amdgpu_kernel void @agent_acquire(
     i32* %in, i32* %out) {
 entry:
@@ -159,12 +159,12 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_seq_cst
-; GCN:       s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
-; GCN-NEXT:  s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  buffer_wbinvl1_vol
-; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
+; GCN-LABEL: {{^}}agent_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}] glc{{$}}
+; GCN-NEXT:   s_waitcnt vmcnt(0){{$}}
+; GFX89-NEXT: buffer_wbinvl1_vol
+; GCN:        flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
 define amdgpu_kernel void @agent_seq_cst(
     i32* %in, i32* %out) {
 entry:
@@ -173,9 +173,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_unordered
+; GCN-LABEL: {{^}}workgroup_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
+; GFX89:     flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:   buffer_wbinvl1_vol
 ; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
@@ -187,9 +187,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_monotonic
+; GCN-LABEL: {{^}}workgroup_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
+; GFX89:     flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN-NOT:   buffer_wbinvl1_vol
 ; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
@@ -201,12 +201,12 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_acquire
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN-NOT:   buffer_wbinvl1_vol
-; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
+; GCN-LABEL: {{^}}workgroup_acquire:
+; GCN-NOT:    s_waitcnt vmcnt(0){{$}}
+; GFX89:      flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
+; GFX89-NOT:  s_waitcnt vmcnt(0){{$}}
+; GFX89-NOT:  buffer_wbinvl1_vol
+; GCN:        flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
 define amdgpu_kernel void @workgroup_acquire(
     i32* %in, i32* %out) {
 entry:
@@ -215,12 +215,12 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_seq_cst
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN-NOT:   buffer_wbinvl1_vol
-; GCN:       flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
+; GCN-LABEL: {{^}}workgroup_seq_cst:
+; GFX89-NOT:  s_waitcnt vmcnt(0){{$}}
+; GFX89:      flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
+; GFX89-NOT:  s_waitcnt vmcnt(0){{$}}
+; GFX89-NOT:  buffer_wbinvl1_vol
+; GCN:        flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[RET]]
 define amdgpu_kernel void @workgroup_seq_cst(
     i32* %in, i32* %out) {
 entry:
@@ -229,7 +229,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_unordered
+; GCN-LABEL: {{^}}wavefront_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -243,7 +243,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_monotonic
+; GCN-LABEL: {{^}}wavefront_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -257,7 +257,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_acquire
+; GCN-LABEL: {{^}}wavefront_acquire:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -271,7 +271,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_seq_cst
+; GCN-LABEL: {{^}}wavefront_seq_cst:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_load_dword [[RET:v[0-9]+]], v[{{[0-9]+}}:{{[0-9]+}}]{{$}}
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
@@ -285,8 +285,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_private_0
-; GCN: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_private_0:
+; GFX89: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
 define amdgpu_kernel void @nontemporal_private_0(
     i32 addrspace(5)* %in, i32* %out) {
 entry:
@@ -295,8 +295,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_private_1
-; GCN: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_private_1:
+; GFX89: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
 define amdgpu_kernel void @nontemporal_private_1(
     i32 addrspace(5)* %in, i32* %out) {
 entry:
@@ -307,7 +307,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_global_0
+; GCN-LABEL: {{^}}nontemporal_global_0:
 ; GCN: s_load_dword s{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], 0x0{{$}}
 define amdgpu_kernel void @nontemporal_global_0(
     i32 addrspace(1)* %in, i32* %out) {
@@ -317,9 +317,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_global_1
-; GFX8: flat_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}] glc slc{{$}}
-; GFX9: global_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}], off glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_global_1:
+; GFX8:  flat_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}] glc slc{{$}}
+; GFX9:  global_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}], off glc slc{{$}}
 define amdgpu_kernel void @nontemporal_global_1(
     i32 addrspace(1)* %in, i32* %out) {
 entry:
@@ -330,7 +330,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_local_0
+; GCN-LABEL: {{^}}nontemporal_local_0:
 ; GCN: ds_read_b32 v{{[0-9]+}}, v{{[0-9]+}}{{$}}
 define amdgpu_kernel void @nontemporal_local_0(
     i32 addrspace(3)* %in, i32* %out) {
@@ -340,7 +340,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_local_1
+; GCN-LABEL: {{^}}nontemporal_local_1:
 ; GCN: ds_read_b32 v{{[0-9]+}}, v{{[0-9]+}}{{$}}
 define amdgpu_kernel void @nontemporal_local_1(
     i32 addrspace(3)* %in, i32* %out) {
@@ -352,8 +352,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_flat_0
-; GCN: flat_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}] glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_flat_0:
+; GFX89: flat_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}] glc slc{{$}}
 define amdgpu_kernel void @nontemporal_flat_0(
     i32* %in, i32* %out) {
 entry:
@@ -362,8 +362,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_flat_1
-; GCN: flat_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}] glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_flat_1:
+; GFX89: flat_load_dword v{{[0-9]+}}, v[{{[0-9]+}}:{{[0-9]+}}] glc slc{{$}}
 define amdgpu_kernel void @nontemporal_flat_1(
     i32* %in, i32* %out) {
 entry:
diff --git a/llvm/test/CodeGen/AMDGPU/memory-legalizer-store.ll b/llvm/test/CodeGen/AMDGPU/memory-legalizer-store.ll
index 296ef1623f3..7bec74d36c8 100644
--- a/llvm/test/CodeGen/AMDGPU/memory-legalizer-store.ll
+++ b/llvm/test/CodeGen/AMDGPU/memory-legalizer-store.ll
@@ -1,11 +1,11 @@
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX8 %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX8 %s
-; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX8,GFX89 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX8,GFX89 %s
+; RUN: llc -mtriple=amdgcn-amd- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX9,GFX89 %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GFX9,GFX89 %s
 
 declare i32 @llvm.amdgcn.workitem.id.x()
 
-; GCN-LABEL: {{^}}system_unordered
+; GCN-LABEL: {{^}}system_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @system_unordered(
@@ -15,7 +15,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}system_monotonic
+; GCN-LABEL: {{^}}system_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @system_monotonic(
@@ -25,9 +25,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}system_release
-; GCN:       s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
+; GCN-LABEL: {{^}}system_release:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @system_release(
     i32 %in, i32* %out) {
 entry:
@@ -35,9 +35,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}system_seq_cst
-; GCN:       s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
+; GCN-LABEL: {{^}}system_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @system_seq_cst(
     i32 %in, i32* %out) {
 entry:
@@ -45,7 +45,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_unordered
+; GCN-LABEL: {{^}}singlethread_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @singlethread_unordered(
@@ -55,7 +55,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_monotonic
+; GCN-LABEL: {{^}}singlethread_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @singlethread_monotonic(
@@ -65,7 +65,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_release
+; GCN-LABEL: {{^}}singlethread_release:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @singlethread_release(
@@ -75,7 +75,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}singlethread_seq_cst
+; GCN-LABEL: {{^}}singlethread_seq_cst:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @singlethread_seq_cst(
@@ -85,7 +85,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_unordered
+; GCN-LABEL: {{^}}agent_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @agent_unordered(
@@ -95,7 +95,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_monotonic
+; GCN-LABEL: {{^}}agent_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @agent_monotonic(
@@ -105,9 +105,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_release
-; GCN:       s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
+; GCN-LABEL: {{^}}agent_release:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @agent_release(
     i32 %in, i32* %out) {
 entry:
@@ -115,9 +115,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}agent_seq_cst
-; GCN:       s_waitcnt vmcnt(0){{$}}
-; GCN-NEXT:  flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
+; GCN-LABEL: {{^}}agent_seq_cst:
+; GCN:        s_waitcnt vmcnt(0){{$}}
+; GCN-NEXT:   flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @agent_seq_cst(
     i32 %in, i32* %out) {
 entry:
@@ -125,7 +125,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_unordered
+; GCN-LABEL: {{^}}workgroup_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @workgroup_unordered(
@@ -135,7 +135,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_monotonic
+; GCN-LABEL: {{^}}workgroup_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @workgroup_monotonic(
@@ -145,9 +145,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_release
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
+; GCN-LABEL: {{^}}workgroup_release:
+; GFX89-NOT:  s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @workgroup_release(
     i32 %in, i32* %out) {
 entry:
@@ -155,9 +155,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}workgroup_seq_cst
-; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
-; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
+; GCN-LABEL: {{^}}workgroup_seq_cst:
+; GFX89-NOT:  s_waitcnt vmcnt(0){{$}}
+; GCN:        flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @workgroup_seq_cst(
     i32 %in, i32* %out) {
 entry:
@@ -165,7 +165,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_unordered
+; GCN-LABEL: {{^}}wavefront_unordered:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @wavefront_unordered(
@@ -175,7 +175,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_monotonic
+; GCN-LABEL: {{^}}wavefront_monotonic:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @wavefront_monotonic(
@@ -185,7 +185,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_release
+; GCN-LABEL: {{^}}wavefront_release:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @wavefront_release(
@@ -195,7 +195,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}wavefront_seq_cst
+; GCN-LABEL: {{^}}wavefront_seq_cst:
 ; GCN-NOT:   s_waitcnt vmcnt(0){{$}}
 ; GCN:       flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], {{v[0-9]+}}{{$}}
 define amdgpu_kernel void @wavefront_seq_cst(
@@ -205,8 +205,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_private_0
-; GCN: buffer_store_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_private_0:
+; GFX89: buffer_store_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
 define amdgpu_kernel void @nontemporal_private_0(
     i32* %in, i32 addrspace(5)* %out) {
 entry:
@@ -215,8 +215,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_private_1
-; GCN: buffer_store_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_private_1:
+; GFX89: buffer_store_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen glc slc{{$}}
 define amdgpu_kernel void @nontemporal_private_1(
     i32* %in, i32 addrspace(5)* %out) {
 entry:
@@ -227,9 +227,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_global_0
-; GFX8: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
-; GFX9: global_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}, off glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_global_0:
+; GFX8:  flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
+; GFX9:  global_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}, off glc slc{{$}}
 define amdgpu_kernel void @nontemporal_global_0(
     i32* %in, i32 addrspace(1)* %out) {
 entry:
@@ -238,9 +238,9 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_global_1
-; GFX8: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
-; GFX9: global_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}, off glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_global_1:
+; GFX8:  flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
+; GFX9:  global_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}}, off glc slc{{$}}
 define amdgpu_kernel void @nontemporal_global_1(
     i32* %in, i32 addrspace(1)* %out) {
 entry:
@@ -251,7 +251,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_local_0
+; GCN-LABEL: {{^}}nontemporal_local_0:
 ; GCN: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}}{{$}}
 define amdgpu_kernel void @nontemporal_local_0(
     i32* %in, i32 addrspace(3)* %out) {
@@ -261,7 +261,7 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_local_1
+; GCN-LABEL: {{^}}nontemporal_local_1:
 ; GCN: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}}{{$}}
 define amdgpu_kernel void @nontemporal_local_1(
     i32* %in, i32 addrspace(3)* %out) {
@@ -273,8 +273,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_flat_0
-; GCN: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_flat_0:
+; GFX89: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
 define amdgpu_kernel void @nontemporal_flat_0(
     i32* %in, i32* %out) {
 entry:
@@ -283,8 +283,8 @@ entry:
   ret void
 }
 
-; GCN-LABEL: {{^}}nontemporal_flat_1
-; GCN: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
+; GCN-LABEL: {{^}}nontemporal_flat_1:
+; GFX89: flat_store_dword v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} glc slc{{$}}
 define amdgpu_kernel void @nontemporal_flat_1(
     i32* %in, i32* %out) {
 entry: