summaryrefslogtreecommitdiff
path: root/llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll
diff options
context:
space:
mode:
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll')
-rw-r--r--llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll362
1 files changed, 272 insertions, 90 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll b/llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll
index d8e6ad043e06..e7f348e9f7d2 100644
--- a/llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll
+++ b/llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-workgroup.ll
@@ -197,7 +197,7 @@ define amdgpu_kernel void @flat_workgroup_unordered_load(
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -390,7 +390,7 @@ define amdgpu_kernel void @flat_workgroup_monotonic_load(
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -595,7 +595,7 @@ define amdgpu_kernel void @flat_workgroup_acquire_load(
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -816,11 +816,12 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_load(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_load_b64 s[2:3], s[4:5], 0x0
; GFX1250-NEXT: s_load_b64 s[0:1], s[4:5], 0x8
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -980,7 +981,8 @@ define amdgpu_kernel void @flat_workgroup_unordered_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -1139,7 +1141,8 @@ define amdgpu_kernel void @flat_workgroup_monotonic_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -1315,8 +1318,10 @@ define amdgpu_kernel void @flat_workgroup_release_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: s_wait_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -1492,8 +1497,10 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: s_wait_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -1652,6 +1659,7 @@ define amdgpu_kernel void @flat_workgroup_monotonic_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
@@ -1830,8 +1838,9 @@ define amdgpu_kernel void @flat_workgroup_acquire_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -2007,7 +2016,9 @@ define amdgpu_kernel void @flat_workgroup_release_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
@@ -2203,9 +2214,11 @@ define amdgpu_kernel void @flat_workgroup_acq_rel_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -2400,9 +2413,11 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -2617,9 +2632,10 @@ define amdgpu_kernel void @flat_workgroup_acquire_ret_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v1, v0, v1, s[0:1] th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -2854,10 +2870,12 @@ define amdgpu_kernel void @flat_workgroup_acq_rel_ret_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v1, v0, v1, s[0:1] th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -3092,10 +3110,12 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_ret_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v1, v0, v1, s[0:1] th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -3347,6 +3367,7 @@ define amdgpu_kernel void @flat_workgroup_monotonic_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
@@ -3618,8 +3639,9 @@ define amdgpu_kernel void @flat_workgroup_acquire_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -3888,7 +3910,9 @@ define amdgpu_kernel void @flat_workgroup_release_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
@@ -4177,9 +4201,11 @@ define amdgpu_kernel void @flat_workgroup_acq_rel_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -4467,9 +4493,11 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -4740,8 +4768,9 @@ define amdgpu_kernel void @flat_workgroup_monotonic_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -5012,8 +5041,9 @@ define amdgpu_kernel void @flat_workgroup_acquire_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -5301,9 +5331,11 @@ define amdgpu_kernel void @flat_workgroup_release_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -5591,9 +5623,11 @@ define amdgpu_kernel void @flat_workgroup_acq_rel_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -5881,9 +5915,11 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -6171,9 +6207,11 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_seq_cst_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt_dscnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -6469,9 +6507,10 @@ define amdgpu_kernel void @flat_workgroup_monotonic_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -6781,9 +6820,10 @@ define amdgpu_kernel void @flat_workgroup_acquire_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -7098,10 +7138,12 @@ define amdgpu_kernel void @flat_workgroup_release_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -7430,10 +7472,12 @@ define amdgpu_kernel void @flat_workgroup_acq_rel_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -7762,10 +7806,12 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -8077,9 +8123,10 @@ define amdgpu_kernel void @flat_workgroup_monotonic_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -8389,9 +8436,10 @@ define amdgpu_kernel void @flat_workgroup_acquire_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -8720,10 +8768,12 @@ define amdgpu_kernel void @flat_workgroup_release_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -9052,10 +9102,12 @@ define amdgpu_kernel void @flat_workgroup_acq_rel_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -9384,10 +9436,12 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -9716,10 +9770,12 @@ define amdgpu_kernel void @flat_workgroup_monotonic_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -10046,10 +10102,12 @@ define amdgpu_kernel void @flat_workgroup_acquire_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -10378,10 +10436,12 @@ define amdgpu_kernel void @flat_workgroup_release_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -10710,10 +10770,12 @@ define amdgpu_kernel void @flat_workgroup_acq_rel_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -11042,10 +11104,12 @@ define amdgpu_kernel void @flat_workgroup_seq_cst_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
-; GFX1250-NEXT: s_wait_dscnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -11240,7 +11304,7 @@ define amdgpu_kernel void @flat_workgroup_one_as_unordered_load(
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -11433,7 +11497,7 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_load(
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -11634,7 +11698,7 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_load(
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -11844,10 +11908,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_load(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_load_b64 s[2:3], s[4:5], 0x0
; GFX1250-NEXT: s_load_b64 s[0:1], s[4:5], 0x8
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: flat_load_b32 v1, v0, s[2:3]
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %in, ptr %out) {
entry:
@@ -12007,7 +12073,8 @@ define amdgpu_kernel void @flat_workgroup_one_as_unordered_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -12166,7 +12233,8 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -12335,7 +12403,10 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -12504,7 +12575,10 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_store(
; GFX1250-NEXT: v_mov_b32_e32 v0, 0
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
i32 %in, ptr %out) {
entry:
@@ -12663,6 +12737,7 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
@@ -12832,7 +12907,9 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -13001,6 +13078,9 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
@@ -13180,7 +13260,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -13359,7 +13443,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v0, v1, s[0:1]
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -13570,9 +13658,10 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_ret_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v1, v0, v1, s[0:1] th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -13796,9 +13885,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_ret_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v1, v0, v1, s[0:1] th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -14022,9 +14114,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_ret_atomicrmw(
; GFX1250-NEXT: s_load_b32 s2, s[4:5], 0x8
; GFX1250-NEXT: s_wait_kmcnt 0x0
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_swap_b32 v1, v0, v1, s[0:1] th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in) {
entry:
@@ -14276,6 +14371,7 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
@@ -14538,7 +14634,9 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -14800,6 +14898,9 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
@@ -15072,7 +15173,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -15344,7 +15449,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_monotonic_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -15606,7 +15715,9 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -15868,7 +15979,9 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -16140,7 +16253,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -16412,7 +16529,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -16684,7 +16805,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_acquire_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -16956,7 +17081,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_seq_cst_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -17228,7 +17357,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_seq_cst_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -17500,7 +17633,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_seq_cst_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -17772,7 +17909,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_seq_cst_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -18044,7 +18185,11 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_seq_cst_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v0, v[2:3], s[0:1] offset:16
+; GFX1250-NEXT: s_wait_storecnt 0x0
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -18340,9 +18485,10 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonicmonotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -18648,9 +18794,10 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -18958,9 +19105,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -19278,9 +19428,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -19598,9 +19751,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_monotonic_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -19908,9 +20064,10 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -20216,9 +20373,10 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -20536,9 +20694,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -20856,9 +21017,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -21176,9 +21340,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_acquire_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -21496,9 +21663,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_monotonic_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -21814,9 +21984,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_acquire_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -22134,9 +22307,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_release_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -22454,9 +22630,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_acq_rel_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
@@ -22774,9 +22953,12 @@ define amdgpu_kernel void @flat_workgroup_one_as_seq_cst_seq_cst_ret_cmpxchg(
; GFX1250-NEXT: v_mov_b32_e32 v1, s2
; GFX1250-NEXT: ; kill: def $vgpr2 killed $vgpr2 def $vgpr2_vgpr3 killed $exec
; GFX1250-NEXT: v_mov_b32_e32 v3, v1
+; GFX1250-NEXT: s_wait_loadcnt 0x0
+; GFX1250-NEXT: s_wait_storecnt 0x0
+; GFX1250-NEXT: s_wait_xcnt 0x0
; GFX1250-NEXT: flat_atomic_cmpswap_b32 v1, v0, v[2:3], s[0:1] offset:16 th:TH_ATOMIC_RETURN
; GFX1250-NEXT: s_wait_loadcnt_dscnt 0x0
-; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1] scope:SCOPE_SE
+; GFX1250-NEXT: flat_store_b32 v0, v1, s[0:1]
; GFX1250-NEXT: s_endpgm
ptr %out, i32 %in, i32 %old) {
entry:
generated by cgit v1.2.3 (git 2.25.1) at 2026-05-31 14:53:30 +0000