summaryrefslogtreecommitdiff
path: root/test/CodeGen/AMDGPU
diff options
context:
space:
mode:
Diffstat (limited to 'test/CodeGen/AMDGPU')
-rw-r--r--test/CodeGen/AMDGPU/GlobalISel/inst-select-load-flat.mir2
-rw-r--r--test/CodeGen/AMDGPU/GlobalISel/inst-select-store-flat.mir2
-rw-r--r--test/CodeGen/AMDGPU/GlobalISel/legalize-constant.mir20
-rw-r--r--test/CodeGen/AMDGPU/GlobalISel/lit.local.cfg2
-rw-r--r--test/CodeGen/AMDGPU/constant-fold-imm-immreg.mir70
-rw-r--r--test/CodeGen/AMDGPU/constant-fold-mi-operands.ll12
-rw-r--r--test/CodeGen/AMDGPU/ctpop.ll80
-rw-r--r--test/CodeGen/AMDGPU/ctpop64.ll16
-rw-r--r--test/CodeGen/AMDGPU/fneg-combines.ll9
-rw-r--r--test/CodeGen/AMDGPU/fneg.f16.ll39
-rw-r--r--test/CodeGen/AMDGPU/inserted-wait-states.mir10
-rw-r--r--test/CodeGen/AMDGPU/limit-coalesce.mir6
-rw-r--r--test/CodeGen/AMDGPU/llvm.amdgcn.cvt.pkrtz.ll18
-rw-r--r--test/CodeGen/AMDGPU/llvm.amdgcn.mbcnt.ll4
-rw-r--r--test/CodeGen/AMDGPU/madak.ll6
-rw-r--r--test/CodeGen/AMDGPU/promote-alloca-volatile.ll12
-rw-r--r--test/CodeGen/AMDGPU/v_madak_f16.ll2
-rw-r--r--test/CodeGen/AMDGPU/waitcnt.mir22
18 files changed, 194 insertions, 138 deletions
diff --git a/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-flat.mir b/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-flat.mir
index 56a9e7022db9c..2a3d3887ed69c 100644
--- a/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-flat.mir
+++ b/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-flat.mir
@@ -14,7 +14,7 @@ regBankSelected: true
# GCN: global_addrspace
# GCN: [[PTR:%[0-9]+]] = COPY %vgpr0_vgpr1
-# GCN: FLAT_LOAD_DWORD [[PTR]], 0, 0, 0
+# GCN: FLAT_LOAD_DWORD [[PTR]], 0, 0
body: |
bb.0:
diff --git a/test/CodeGen/AMDGPU/GlobalISel/inst-select-store-flat.mir b/test/CodeGen/AMDGPU/GlobalISel/inst-select-store-flat.mir
index ea435725bf25d..89be3bde94a8d 100644
--- a/test/CodeGen/AMDGPU/GlobalISel/inst-select-store-flat.mir
+++ b/test/CodeGen/AMDGPU/GlobalISel/inst-select-store-flat.mir
@@ -15,7 +15,7 @@ regBankSelected: true
# GCN: global_addrspace
# GCN: [[PTR:%[0-9]+]] = COPY %vgpr0_vgpr1
# GCN: [[VAL:%[0-9]+]] = COPY %vgpr2
-# GCN: FLAT_STORE_DWORD [[PTR]], [[VAL]], 0, 0, 0
+# GCN: FLAT_STORE_DWORD [[PTR]], [[VAL]], 0, 0
body: |
bb.0:
diff --git a/test/CodeGen/AMDGPU/GlobalISel/legalize-constant.mir b/test/CodeGen/AMDGPU/GlobalISel/legalize-constant.mir
new file mode 100644
index 0000000000000..8839ba8e0ab24
--- /dev/null
+++ b/test/CodeGen/AMDGPU/GlobalISel/legalize-constant.mir
@@ -0,0 +1,20 @@
+# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -O0 -run-pass=legalizer -global-isel %s -o - | FileCheck %s
+
+--- |
+ define void @test_constant() {
+ entry:
+ ret void
+ }
+...
+
+---
+name: test_constant
+registers:
+ - { id: 0, class: _ }
+body: |
+ bb.0.entry:
+ ; CHECK-LABEL: name: test_constant
+ ; CHECK: %0(s32) = G_CONSTANT i32 5
+
+ %0(s32) = G_CONSTANT i32 5
+...
diff --git a/test/CodeGen/AMDGPU/GlobalISel/lit.local.cfg b/test/CodeGen/AMDGPU/GlobalISel/lit.local.cfg
new file mode 100644
index 0000000000000..e99d1bb8446ce
--- /dev/null
+++ b/test/CodeGen/AMDGPU/GlobalISel/lit.local.cfg
@@ -0,0 +1,2 @@
+if not 'global-isel' in config.root.available_features:
+ config.unsupported = True
diff --git a/test/CodeGen/AMDGPU/constant-fold-imm-immreg.mir b/test/CodeGen/AMDGPU/constant-fold-imm-immreg.mir
index 62b47beb12518..bc992ed77ffdb 100644
--- a/test/CodeGen/AMDGPU/constant-fold-imm-immreg.mir
+++ b/test/CodeGen/AMDGPU/constant-fold-imm-immreg.mir
@@ -219,19 +219,19 @@ body: |
%34 = V_MOV_B32_e32 63, implicit %exec
%27 = V_AND_B32_e64 %26, %24, implicit %exec
- FLAT_STORE_DWORD %37, %27, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %37, %27, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%28 = V_AND_B32_e64 %24, %26, implicit %exec
- FLAT_STORE_DWORD %37, %28, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %37, %28, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%29 = V_AND_B32_e32 %26, %24, implicit %exec
- FLAT_STORE_DWORD %37, %29, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %37, %29, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%30 = V_AND_B32_e64 %26, %26, implicit %exec
- FLAT_STORE_DWORD %37, %30, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %37, %30, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%31 = V_AND_B32_e64 %34, %34, implicit %exec
- FLAT_STORE_DWORD %37, %31, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %37, %31, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
S_ENDPGM
@@ -407,34 +407,34 @@ body: |
%27 = S_MOV_B32 -4
%11 = V_LSHLREV_B32_e64 12, %10, implicit %exec
- FLAT_STORE_DWORD %20, %11, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %11, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%12 = V_LSHLREV_B32_e64 %7, 12, implicit %exec
- FLAT_STORE_DWORD %20, %12, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %12, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%13 = V_LSHL_B32_e64 %7, 12, implicit %exec
- FLAT_STORE_DWORD %20, %13, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %13, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%14 = V_LSHL_B32_e64 12, %7, implicit %exec
- FLAT_STORE_DWORD %20, %14, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %14, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%15 = V_LSHL_B32_e64 12, %24, implicit %exec
- FLAT_STORE_DWORD %20, %15, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %15, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%22 = V_LSHL_B32_e64 %6, 12, implicit %exec
- FLAT_STORE_DWORD %20, %22, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %22, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%23 = V_LSHL_B32_e64 %6, 32, implicit %exec
- FLAT_STORE_DWORD %20, %23, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %23, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%25 = V_LSHL_B32_e32 %6, %6, implicit %exec
- FLAT_STORE_DWORD %20, %25, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %25, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%26 = V_LSHLREV_B32_e32 11, %24, implicit %exec
- FLAT_STORE_DWORD %20, %26, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %26, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%28 = V_LSHL_B32_e32 %27, %6, implicit %exec
- FLAT_STORE_DWORD %20, %28, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %28, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
S_ENDPGM
@@ -615,34 +615,34 @@ body: |
%35 = V_MOV_B32_e32 2, implicit %exec
%11 = V_ASHRREV_I32_e64 8, %10, implicit %exec
- FLAT_STORE_DWORD %20, %11, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %11, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%12 = V_ASHRREV_I32_e64 %8, %10, implicit %exec
- FLAT_STORE_DWORD %20, %12, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %12, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%13 = V_ASHR_I32_e64 %7, 3, implicit %exec
- FLAT_STORE_DWORD %20, %13, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %13, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%14 = V_ASHR_I32_e64 7, %32, implicit %exec
- FLAT_STORE_DWORD %20, %14, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %14, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%15 = V_ASHR_I32_e64 %27, %24, implicit %exec
- FLAT_STORE_DWORD %20, %15, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %15, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%22 = V_ASHR_I32_e64 %6, 4, implicit %exec
- FLAT_STORE_DWORD %20, %22, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %22, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%23 = V_ASHR_I32_e64 %6, %33, implicit %exec
- FLAT_STORE_DWORD %20, %23, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %23, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%25 = V_ASHR_I32_e32 %34, %34, implicit %exec
- FLAT_STORE_DWORD %20, %25, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %25, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%26 = V_ASHRREV_I32_e32 11, %10, implicit %exec
- FLAT_STORE_DWORD %20, %26, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %26, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%28 = V_ASHR_I32_e32 %27, %35, implicit %exec
- FLAT_STORE_DWORD %20, %28, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %28, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
S_ENDPGM
@@ -824,34 +824,34 @@ body: |
%35 = V_MOV_B32_e32 2, implicit %exec
%11 = V_LSHRREV_B32_e64 8, %10, implicit %exec
- FLAT_STORE_DWORD %20, %11, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %11, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%12 = V_LSHRREV_B32_e64 %8, %10, implicit %exec
- FLAT_STORE_DWORD %20, %12, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %12, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%13 = V_LSHR_B32_e64 %7, 3, implicit %exec
- FLAT_STORE_DWORD %20, %13, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %13, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%14 = V_LSHR_B32_e64 7, %32, implicit %exec
- FLAT_STORE_DWORD %20, %14, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %14, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%15 = V_LSHR_B32_e64 %27, %24, implicit %exec
- FLAT_STORE_DWORD %20, %15, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %15, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%22 = V_LSHR_B32_e64 %6, 4, implicit %exec
- FLAT_STORE_DWORD %20, %22, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %22, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%23 = V_LSHR_B32_e64 %6, %33, implicit %exec
- FLAT_STORE_DWORD %20, %23, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %23, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%25 = V_LSHR_B32_e32 %34, %34, implicit %exec
- FLAT_STORE_DWORD %20, %25, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %25, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%26 = V_LSHRREV_B32_e32 11, %10, implicit %exec
- FLAT_STORE_DWORD %20, %26, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %26, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
%28 = V_LSHR_B32_e32 %27, %35, implicit %exec
- FLAT_STORE_DWORD %20, %28, 0, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
+ FLAT_STORE_DWORD %20, %28, 0, 0, implicit %exec, implicit %flat_scr :: (volatile store 4 into %ir.gep.out)
S_ENDPGM
diff --git a/test/CodeGen/AMDGPU/constant-fold-mi-operands.ll b/test/CodeGen/AMDGPU/constant-fold-mi-operands.ll
index 0831d250b9e7d..8611cd080e15d 100644
--- a/test/CodeGen/AMDGPU/constant-fold-mi-operands.ll
+++ b/test/CodeGen/AMDGPU/constant-fold-mi-operands.ll
@@ -25,7 +25,7 @@ define amdgpu_kernel void @fold_mi_s_and_0(i32 addrspace(1)* %out, i32 %x) #0 {
}
; GCN-LABEL: {{^}}fold_mi_v_or_0:
-; GCN: v_mbcnt_lo_u32_b32_e64 [[RESULT:v[0-9]+]]
+; GCN: v_mbcnt_lo_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]]
; GCN-NOT: [[RESULT]]
; GCN: buffer_store_dword [[RESULT]]
define amdgpu_kernel void @fold_mi_v_or_0(i32 addrspace(1)* %out) {
@@ -50,7 +50,7 @@ define amdgpu_kernel void @fold_mi_s_or_0(i32 addrspace(1)* %out, i32 %x) #0 {
}
; GCN-LABEL: {{^}}fold_mi_v_xor_0:
-; GCN: v_mbcnt_lo_u32_b32_e64 [[RESULT:v[0-9]+]]
+; GCN: v_mbcnt_lo_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]]
; GCN-NOT: [[RESULT]]
; GCN: buffer_store_dword [[RESULT]]
define amdgpu_kernel void @fold_mi_v_xor_0(i32 addrspace(1)* %out) {
@@ -86,8 +86,8 @@ define amdgpu_kernel void @fold_mi_s_not_0(i32 addrspace(1)* %out, i32 %x) #0 {
}
; GCN-LABEL: {{^}}fold_mi_v_not_0:
-; GCN: v_bcnt_u32_b32_e64 v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, 0{{$}}
-; GCN: v_bcnt_u32_b32_e{{[0-9]+}} v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, v[[RESULT_LO]]{{$}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}} v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, 0{{$}}
+; GCN: v_bcnt_u32_b32{{(_e32)*(_e64)*}} v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, v[[RESULT_LO]]{{$}}
; GCN-NEXT: v_not_b32_e32 v[[RESULT_LO]]
; GCN-NEXT: v_mov_b32_e32 v[[RESULT_HI:[0-9]+]], -1{{$}}
; GCN-NEXT: buffer_store_dwordx2 v{{\[}}[[RESULT_LO]]:[[RESULT_HI]]{{\]}}
@@ -104,8 +104,8 @@ define amdgpu_kernel void @fold_mi_v_not_0(i64 addrspace(1)* %out) {
; GCN: buffer_load_dwordx2
; GCN: buffer_load_dwordx2 v{{\[}}[[VREG1_LO:[0-9]+]]:[[VREG1_HI:[0-9]+]]{{\]}}
-; GCN: v_bcnt_u32_b32_e64 v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, 0{{$}}
-; GCN: v_bcnt_u32_b32_e{{[0-9]+}} v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, v[[RESULT_LO]]{{$}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}} v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, 0{{$}}
+; GCN: v_bcnt_u32_b32{{(_e32)*(_e64)*}} v[[RESULT_LO:[0-9]+]], v{{[0-9]+}}, v[[RESULT_LO]]{{$}}
; GCN-DAG: v_not_b32_e32 v[[RESULT_LO]], v[[RESULT_LO]]
; GCN-DAG: v_or_b32_e32 v[[RESULT_LO]], v[[VREG1_LO]], v[[RESULT_LO]]
; GCN-DAG: v_mov_b32_e32 v[[RESULT_HI:[0-9]+]], v[[VREG1_HI]]
diff --git a/test/CodeGen/AMDGPU/ctpop.ll b/test/CodeGen/AMDGPU/ctpop.ll
index a29e72ea57cb3..aa913ad406d2b 100644
--- a/test/CodeGen/AMDGPU/ctpop.ll
+++ b/test/CodeGen/AMDGPU/ctpop.ll
@@ -25,7 +25,7 @@ define amdgpu_kernel void @s_ctpop_i32(i32 addrspace(1)* noalias %out, i32 %val)
; XXX - Why 0 in register?
; FUNC-LABEL: {{^}}v_ctpop_i32:
; GCN: buffer_load_dword [[VAL:v[0-9]+]],
-; GCN: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], 0
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]], [[VAL]], 0
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
@@ -40,9 +40,9 @@ define amdgpu_kernel void @v_ctpop_i32(i32 addrspace(1)* noalias %out, i32 addrs
; FUNC-LABEL: {{^}}v_ctpop_add_chain_i32:
; GCN: buffer_load_dword [[VAL1:v[0-9]+]],
; GCN: buffer_load_dword [[VAL0:v[0-9]+]],
-; GCN: v_bcnt_u32_b32_e64 [[MIDRESULT:v[0-9]+]], [[VAL1]], 0
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT:v[0-9]+]], [[VAL1]], 0
; SI: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], [[VAL0]], [[MIDRESULT]]
-; VI: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL0]], [[MIDRESULT]]
+; VI: v_bcnt_u32_b32 [[RESULT:v[0-9]+]], [[VAL0]], [[MIDRESULT]]
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
@@ -61,7 +61,7 @@ define amdgpu_kernel void @v_ctpop_add_chain_i32(i32 addrspace(1)* noalias %out,
; FUNC-LABEL: {{^}}v_ctpop_add_sgpr_i32:
; GCN: buffer_load_dword [[VAL0:v[0-9]+]],
; GCN: s_waitcnt
-; GCN-NEXT: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL0]], s{{[0-9]+}}
+; GCN-NEXT: v_bcnt_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]], [[VAL0]], s{{[0-9]+}}
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_add_sgpr_i32(i32 addrspace(1)* noalias %out, i32 addrspace(1)* noalias %in0, i32 addrspace(1)* noalias %in1, i32 %sval) nounwind {
@@ -73,8 +73,8 @@ define amdgpu_kernel void @v_ctpop_add_sgpr_i32(i32 addrspace(1)* noalias %out,
}
; FUNC-LABEL: {{^}}v_ctpop_v2i32:
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
; GCN: s_endpgm
; EG: BCNT_INT
@@ -87,10 +87,10 @@ define amdgpu_kernel void @v_ctpop_v2i32(<2 x i32> addrspace(1)* noalias %out, <
}
; FUNC-LABEL: {{^}}v_ctpop_v4i32:
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
; GCN: s_endpgm
; EG: BCNT_INT
@@ -105,14 +105,14 @@ define amdgpu_kernel void @v_ctpop_v4i32(<4 x i32> addrspace(1)* noalias %out, <
}
; FUNC-LABEL: {{^}}v_ctpop_v8i32:
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
; GCN: s_endpgm
; EG: BCNT_INT
@@ -131,22 +131,22 @@ define amdgpu_kernel void @v_ctpop_v8i32(<8 x i32> addrspace(1)* noalias %out, <
}
; FUNC-LABEL: {{^}}v_ctpop_v16i32:
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
-; GCN: v_bcnt_u32_b32_e64
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
+; GCN: v_bcnt_u32_b32{{(_e64)*}}
; GCN: s_endpgm
; EG: BCNT_INT
@@ -174,7 +174,7 @@ define amdgpu_kernel void @v_ctpop_v16i32(<16 x i32> addrspace(1)* noalias %out,
; FUNC-LABEL: {{^}}v_ctpop_i32_add_inline_constant:
; GCN: buffer_load_dword [[VAL:v[0-9]+]],
-; GCN: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], 4
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]], [[VAL]], 4
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
@@ -189,7 +189,7 @@ define amdgpu_kernel void @v_ctpop_i32_add_inline_constant(i32 addrspace(1)* noa
; FUNC-LABEL: {{^}}v_ctpop_i32_add_inline_constant_inv:
; GCN: buffer_load_dword [[VAL:v[0-9]+]],
-; GCN: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], 4
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]], [[VAL]], 4
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
@@ -206,7 +206,7 @@ define amdgpu_kernel void @v_ctpop_i32_add_inline_constant_inv(i32 addrspace(1)*
; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]],
; GCN-DAG: v_mov_b32_e32 [[LIT:v[0-9]+]], 0x1869f
; SI: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], [[VAL]], [[LIT]]
-; VI: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], [[LIT]]
+; VI: v_bcnt_u32_b32 [[RESULT:v[0-9]+]], [[VAL]], [[LIT]]
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_i32_add_literal(i32 addrspace(1)* noalias %out, i32 addrspace(1)* noalias %in) nounwind {
@@ -220,7 +220,7 @@ define amdgpu_kernel void @v_ctpop_i32_add_literal(i32 addrspace(1)* noalias %ou
; FUNC-LABEL: {{^}}v_ctpop_i32_add_var:
; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]],
; GCN-DAG: s_load_dword [[VAR:s[0-9]+]],
-; GCN: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
@@ -236,7 +236,7 @@ define amdgpu_kernel void @v_ctpop_i32_add_var(i32 addrspace(1)* noalias %out, i
; FUNC-LABEL: {{^}}v_ctpop_i32_add_var_inv:
; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]],
; GCN-DAG: s_load_dword [[VAR:s[0-9]+]],
-; GCN: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
@@ -253,7 +253,7 @@ define amdgpu_kernel void @v_ctpop_i32_add_var_inv(i32 addrspace(1)* noalias %ou
; GCN-DAG: buffer_load_dword [[VAL:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], {{0$}}
; GCN-DAG: buffer_load_dword [[VAR:v[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0 offset:16
; SI: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
-; VI: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
+; VI: v_bcnt_u32_b32 [[RESULT:v[0-9]+]], [[VAL]], [[VAR]]
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
diff --git a/test/CodeGen/AMDGPU/ctpop64.ll b/test/CodeGen/AMDGPU/ctpop64.ll
index 2610684ad9ee7..f18bd9fd8174b 100644
--- a/test/CodeGen/AMDGPU/ctpop64.ll
+++ b/test/CodeGen/AMDGPU/ctpop64.ll
@@ -26,9 +26,9 @@ define amdgpu_kernel void @s_ctpop_i64(i32 addrspace(1)* noalias %out, i64 %val)
; FUNC-LABEL: {{^}}v_ctpop_i64:
; GCN: buffer_load_dwordx2 v{{\[}}[[LOVAL:[0-9]+]]:[[HIVAL:[0-9]+]]{{\]}},
-; GCN: v_bcnt_u32_b32_e64 [[MIDRESULT:v[0-9]+]], v[[LOVAL]], 0
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT:v[0-9]+]], v[[LOVAL]], 0
; SI-NEXT: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
-; VI-NEXT: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
+; VI-NEXT: v_bcnt_u32_b32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_i64(i32 addrspace(1)* noalias %out, i64 addrspace(1)* noalias %in) nounwind {
@@ -41,9 +41,9 @@ define amdgpu_kernel void @v_ctpop_i64(i32 addrspace(1)* noalias %out, i64 addrs
; FUNC-LABEL: {{^}}v_ctpop_i64_user:
; GCN: buffer_load_dwordx2 v{{\[}}[[LOVAL:[0-9]+]]:[[HIVAL:[0-9]+]]{{\]}},
-; GCN: v_bcnt_u32_b32_e64 [[MIDRESULT:v[0-9]+]], v[[LOVAL]], 0
+; GCN: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT:v[0-9]+]], v[[LOVAL]], 0
; SI-NEXT: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
-; VI-NEXT: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
+; VI-NEXT: v_bcnt_u32_b32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
; GCN-DAG: v_or_b32_e32 v[[RESULT_LO:[0-9]+]], s{{[0-9]+}}, [[RESULT]]
; GCN-DAG: v_mov_b32_e32 v[[RESULT_HI:[0-9]+]], s{{[0-9]+}}
; GCN: buffer_store_dwordx2 v{{\[}}[[RESULT_LO]]:[[RESULT_HI]]{{\]}}
@@ -171,11 +171,11 @@ define amdgpu_kernel void @s_ctpop_i65(i32 addrspace(1)* noalias %out, i65 %val)
; FUNC-LABEL: {{^}}v_ctpop_i128:
; GCN: buffer_load_dwordx4 v{{\[}}[[VAL0:[0-9]+]]:[[VAL3:[0-9]+]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
-; GCN-DAG: v_bcnt_u32_b32_e64 [[MIDRESULT0:v[0-9]+]], v{{[0-9]+}}, 0
-; GCN-DAG: v_bcnt_u32_b32{{_e32|_e64}} [[MIDRESULT1:v[0-9]+]], v[[VAL3]], [[MIDRESULT0]]
+; GCN-DAG: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT0:v[0-9]+]], v{{[0-9]+}}, 0
+; GCN-DAG: v_bcnt_u32_b32{{(_e32)*(_e64)*}} [[MIDRESULT1:v[0-9]+]], v[[VAL3]], [[MIDRESULT0]]
-; GCN-DAG: v_bcnt_u32_b32_e64 [[MIDRESULT2:v[0-9]+]], v[[VAL0]], 0
-; GCN-DAG: v_bcnt_u32_b32{{_e32|_e64}} [[MIDRESULT3:v[0-9]+]], v{{[0-9]+}}, [[MIDRESULT2]]
+; GCN-DAG: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT2:v[0-9]+]], v[[VAL0]], 0
+; GCN-DAG: v_bcnt_u32_b32{{(_e32)*(_e64)*}} [[MIDRESULT3:v[0-9]+]], v{{[0-9]+}}, [[MIDRESULT2]]
; GCN: v_add_i32_e32 [[RESULT:v[0-9]+]], vcc, [[MIDRESULT1]], [[MIDRESULT2]]
diff --git a/test/CodeGen/AMDGPU/fneg-combines.ll b/test/CodeGen/AMDGPU/fneg-combines.ll
index 1c0e9a2f13cea..66bf9d0ffb00e 100644
--- a/test/CodeGen/AMDGPU/fneg-combines.ll
+++ b/test/CodeGen/AMDGPU/fneg-combines.ll
@@ -1471,11 +1471,10 @@ define amdgpu_kernel void @v_fneg_mul_legacy_store_use_mul_legacy_f32(float addr
; GCN-LABEL: {{^}}v_fneg_mul_legacy_multi_use_mul_legacy_f32:
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[B:v[0-9]+]]
-; GCN-DAG: v_mul_legacy_f32_e32 [[ADD:v[0-9]+]], [[B]], [[A]]
-; GCN-DAG: v_xor_b32_e32 [[NEG_MUL_LEGACY:v[0-9]+]], 0x80000000, [[ADD]]
-; GCN: v_mul_legacy_f32_e32 [[MUL:v[0-9]+]], 4.0, [[ADD]]
-; GCN-NEXT: buffer_store_dword [[NEG_MUL_LEGACY]]
-; GCN: buffer_store_dword [[MUL]]
+; GCN: v_mul_legacy_f32_e64 [[ADD:v[0-9]+]], [[A]], -[[B]]
+; GCN-NEXT: v_mul_legacy_f32_e64 [[MUL:v[0-9]+]], -[[ADD]], 4.0
+; GCN-NEXT: buffer_store_dword [[ADD]]
+; GCN-NEXT: buffer_store_dword [[MUL]]
define amdgpu_kernel void @v_fneg_mul_legacy_multi_use_mul_legacy_f32(float addrspace(1)* %out, float addrspace(1)* %a.ptr, float addrspace(1)* %b.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
diff --git a/test/CodeGen/AMDGPU/fneg.f16.ll b/test/CodeGen/AMDGPU/fneg.f16.ll
index 626a0b50cce8a..ed36666db807d 100644
--- a/test/CodeGen/AMDGPU/fneg.f16.ll
+++ b/test/CodeGen/AMDGPU/fneg.f16.ll
@@ -1,6 +1,6 @@
; RUN: llc -march=amdgcn -mcpu=kaveri -mtriple=amdgcn--amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=CI -check-prefix=CIVI -check-prefix=GCN %s
-; RUN: llc -march=amdgcn -mcpu=tonga -mtriple=amdgcn--amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=VI -check-prefix=CIVI -check-prefix=GCN %s
-; RUN: llc -march=amdgcn -mcpu=gfx901 -mtriple=amdgcn--amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GFX9 -check-prefix=GCN %s
+; RUN: llc -march=amdgcn -mcpu=tonga -mtriple=amdgcn--amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=VI -check-prefix=CIVI -check-prefix=GCN -check-prefix=GFX89 %s
+; RUN: llc -march=amdgcn -mcpu=gfx901 -mtriple=amdgcn--amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GFX9 -check-prefix=GCN -check-prefix=GFX89 %s
; FIXME: Should be able to do scalar op
; GCN-LABEL: {{^}}s_fneg_f16:
@@ -129,6 +129,41 @@ define amdgpu_kernel void @v_fneg_fold_v2f16(<2 x half> addrspace(1)* %out, <2 x
ret void
}
+; GCN-LABEL: {{^}}v_extract_fneg_fold_v2f16:
+; GCN: flat_load_dword [[VAL:v[0-9]+]]
+; CI-DAG: v_mul_f32_e32 v{{[0-9]+}}, -4.0, v{{[0-9]+}}
+; CI-DAG: v_sub_f32_e32 v{{[0-9]+}}, 2.0, v{{[0-9]+}}
+
+; GFX89: v_lshrrev_b32_e32 [[ELT1:v[0-9]+]], 16, [[VAL]]
+; GFX89-DAG: v_mul_f16_e32 v{{[0-9]+}}, -4.0, [[VAL]]
+; GFX89-DAG: v_sub_f16_e32 v{{[0-9]+}}, 2.0, [[ELT1]]
+define amdgpu_kernel void @v_extract_fneg_fold_v2f16(<2 x half> addrspace(1)* %in) #0 {
+ %val = load <2 x half>, <2 x half> addrspace(1)* %in
+ %fneg = fsub <2 x half> <half -0.0, half -0.0>, %val
+ %elt0 = extractelement <2 x half> %fneg, i32 0
+ %elt1 = extractelement <2 x half> %fneg, i32 1
+
+ %fmul0 = fmul half %elt0, 4.0
+ %fadd1 = fadd half %elt1, 2.0
+ store volatile half %fmul0, half addrspace(1)* undef
+ store volatile half %fadd1, half addrspace(1)* undef
+ ret void
+}
+
+; GCN-LABEL: {{^}}v_extract_fneg_no_fold_v2f16:
+; GCN: flat_load_dword [[VAL:v[0-9]+]]
+; GCN: v_xor_b32_e32 [[NEG:v[0-9]+]], 0x80008000, [[VAL]]
+; GCN: v_lshrrev_b32_e32 [[ELT1:v[0-9]+]], 16, [[NEG]]
+define amdgpu_kernel void @v_extract_fneg_no_fold_v2f16(<2 x half> addrspace(1)* %in) #0 {
+ %val = load <2 x half>, <2 x half> addrspace(1)* %in
+ %fneg = fsub <2 x half> <half -0.0, half -0.0>, %val
+ %elt0 = extractelement <2 x half> %fneg, i32 0
+ %elt1 = extractelement <2 x half> %fneg, i32 1
+ store volatile half %elt0, half addrspace(1)* undef
+ store volatile half %elt1, half addrspace(1)* undef
+ ret void
+}
+
declare i32 @llvm.amdgcn.workitem.id.x() #1
attributes #0 = { nounwind }
diff --git a/test/CodeGen/AMDGPU/inserted-wait-states.mir b/test/CodeGen/AMDGPU/inserted-wait-states.mir
index c6fe6debd225a..ff9fcd1c693f7 100644
--- a/test/CodeGen/AMDGPU/inserted-wait-states.mir
+++ b/test/CodeGen/AMDGPU/inserted-wait-states.mir
@@ -246,15 +246,15 @@ body: |
S_BRANCH %bb.1
bb.1:
- FLAT_STORE_DWORDX2 %vgpr0_vgpr1, %vgpr2_vgpr3, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORDX2 %vgpr0_vgpr1, %vgpr2_vgpr3, 0, 0, implicit %exec, implicit %flat_scr
%vgpr3 = V_MOV_B32_e32 0, implicit %exec
- FLAT_STORE_DWORDX3 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORDX3 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4, 0, 0, implicit %exec, implicit %flat_scr
%vgpr3 = V_MOV_B32_e32 0, implicit %exec
- FLAT_STORE_DWORDX4 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4_vgpr5, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORDX4 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4_vgpr5, 0, 0, implicit %exec, implicit %flat_scr
%vgpr3 = V_MOV_B32_e32 0, implicit %exec
- FLAT_ATOMIC_CMPSWAP_X2 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4_vgpr5, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_ATOMIC_CMPSWAP_X2 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4_vgpr5, 0, implicit %exec, implicit %flat_scr
%vgpr3 = V_MOV_B32_e32 0, implicit %exec
- FLAT_ATOMIC_FCMPSWAP_X2 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4_vgpr5, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_ATOMIC_FCMPSWAP_X2 %vgpr0_vgpr1, %vgpr2_vgpr3_vgpr4_vgpr5, 0, implicit %exec, implicit %flat_scr
%vgpr3 = V_MOV_B32_e32 0, implicit %exec
S_ENDPGM
diff --git a/test/CodeGen/AMDGPU/limit-coalesce.mir b/test/CodeGen/AMDGPU/limit-coalesce.mir
index 106a96e32dc38..a0d2d6c097a27 100644
--- a/test/CodeGen/AMDGPU/limit-coalesce.mir
+++ b/test/CodeGen/AMDGPU/limit-coalesce.mir
@@ -57,15 +57,15 @@ body: |
%4.sub1 = COPY %3.sub0
undef %5.sub0 = COPY %4.sub1
%5.sub1 = COPY %4.sub0
- FLAT_STORE_DWORDX2 %vgpr0_vgpr1, killed %5, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORDX2 %vgpr0_vgpr1, killed %5, 0, 0, implicit %exec, implicit %flat_scr
%6 = IMPLICIT_DEF
undef %7.sub0_sub1 = COPY %6
%7.sub2 = COPY %3.sub0
- FLAT_STORE_DWORDX3 %vgpr0_vgpr1, killed %7, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORDX3 %vgpr0_vgpr1, killed %7, 0, 0, implicit %exec, implicit %flat_scr
%8 = IMPLICIT_DEF
undef %9.sub0_sub1_sub2 = COPY %8
%9.sub3 = COPY %3.sub0
- FLAT_STORE_DWORDX4 %vgpr0_vgpr1, killed %9, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORDX4 %vgpr0_vgpr1, killed %9, 0, 0, implicit %exec, implicit %flat_scr
...
diff --git a/test/CodeGen/AMDGPU/llvm.amdgcn.cvt.pkrtz.ll b/test/CodeGen/AMDGPU/llvm.amdgcn.cvt.pkrtz.ll
index b92eb34750d9a..7179d02fc6ddc 100644
--- a/test/CodeGen/AMDGPU/llvm.amdgcn.cvt.pkrtz.ll
+++ b/test/CodeGen/AMDGPU/llvm.amdgcn.cvt.pkrtz.ll
@@ -7,7 +7,7 @@
; GCN-DAG: s_load_dword [[SY:s[0-9]+]], s[0:1], 0x{{c|30}}
; GCN: v_mov_b32_e32 [[VY:v[0-9]+]], [[SY]]
; SI: v_cvt_pkrtz_f16_f32_e32 v{{[0-9]+}}, [[X]], [[VY]]
-; GFX89: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, [[X]], [[VY]]
+; GFX89: v_cvt_pkrtz_f16_f32 v{{[0-9]+}}, [[X]], [[VY]]
define amdgpu_kernel void @s_cvt_pkrtz_v2f16_f32(<2 x half> addrspace(1)* %out, float %x, float %y) #0 {
%result = call <2 x half> @llvm.amdgcn.cvt.pkrtz(float %x, float %y)
store <2 x half> %result, <2 x half> addrspace(1)* %out
@@ -16,7 +16,7 @@ define amdgpu_kernel void @s_cvt_pkrtz_v2f16_f32(<2 x half> addrspace(1)* %out,
; GCN-LABEL: {{^}}s_cvt_pkrtz_samereg_v2f16_f32:
; GCN: s_load_dword [[X:s[0-9]+]]
-; GCN: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, [[X]], [[X]]
+; GCN: v_cvt_pkrtz_f16_f32{{(_e64)*}} v{{[0-9]+}}, [[X]], [[X]]
define amdgpu_kernel void @s_cvt_pkrtz_samereg_v2f16_f32(<2 x half> addrspace(1)* %out, float %x) #0 {
%result = call <2 x half> @llvm.amdgcn.cvt.pkrtz(float %x, float %x)
store <2 x half> %result, <2 x half> addrspace(1)* %out
@@ -39,7 +39,7 @@ define amdgpu_kernel void @s_cvt_pkrtz_undef_undef(<2 x half> addrspace(1)* %out
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[B:v[0-9]+]]
; SI: v_cvt_pkrtz_f16_f32_e32 v{{[0-9]+}}, [[A]], [[B]]
-; GFX89: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, [[A]], [[B]]
+; GFX89: v_cvt_pkrtz_f16_f32 v{{[0-9]+}}, [[A]], [[B]]
define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32(<2 x half> addrspace(1)* %out, float addrspace(1)* %a.ptr, float addrspace(1)* %b.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
@@ -55,7 +55,7 @@ define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32(<2 x half> addrspace(1)* %out,
; GCN-LABEL: {{^}}v_cvt_pkrtz_v2f16_f32_reg_imm:
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
-; GCN: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, [[A]], 1.0
+; GCN: v_cvt_pkrtz_f16_f32{{(_e64)*}} v{{[0-9]+}}, [[A]], 1.0
define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_reg_imm(<2 x half> addrspace(1)* %out, float addrspace(1)* %a.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
@@ -70,7 +70,7 @@ define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_reg_imm(<2 x half> addrspace(1)
; GCN-LABEL: {{^}}v_cvt_pkrtz_v2f16_f32_imm_reg:
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
; SI: v_cvt_pkrtz_f16_f32_e32 v{{[0-9]+}}, 1.0, [[A]]
-; GFX89: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, 1.0, [[A]]
+; GFX89: v_cvt_pkrtz_f16_f32 v{{[0-9]+}}, 1.0, [[A]]
define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_imm_reg(<2 x half> addrspace(1)* %out, float addrspace(1)* %a.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
@@ -85,7 +85,7 @@ define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_imm_reg(<2 x half> addrspace(1)
; GCN-LABEL: {{^}}v_cvt_pkrtz_v2f16_f32_fneg_lo:
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[B:v[0-9]+]]
-; GCN: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, -[[A]], [[B]]
+; GCN: v_cvt_pkrtz_f16_f32{{(_e64)*}} v{{[0-9]+}}, -[[A]], [[B]]
define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_fneg_lo(<2 x half> addrspace(1)* %out, float addrspace(1)* %a.ptr, float addrspace(1)* %b.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
@@ -103,7 +103,7 @@ define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_fneg_lo(<2 x half> addrspace(1)
; GCN-LABEL: {{^}}v_cvt_pkrtz_v2f16_f32_fneg_hi:
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[B:v[0-9]+]]
-; GCN: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, [[A]], -[[B]]
+; GCN: v_cvt_pkrtz_f16_f32{{(_e64)*}} v{{[0-9]+}}, [[A]], -[[B]]
define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_fneg_hi(<2 x half> addrspace(1)* %out, float addrspace(1)* %a.ptr, float addrspace(1)* %b.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
@@ -121,7 +121,7 @@ define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_fneg_hi(<2 x half> addrspace(1)
; GCN-LABEL: {{^}}v_cvt_pkrtz_v2f16_f32_fneg_lo_hi:
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[B:v[0-9]+]]
-; GCN: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, -[[A]], -[[B]]
+; GCN: v_cvt_pkrtz_f16_f32{{(_e64)*}} v{{[0-9]+}}, -[[A]], -[[B]]
define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_fneg_lo_hi(<2 x half> addrspace(1)* %out, float addrspace(1)* %a.ptr, float addrspace(1)* %b.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
@@ -140,7 +140,7 @@ define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_fneg_lo_hi(<2 x half> addrspace
; GCN-LABEL: {{^}}v_cvt_pkrtz_v2f16_f32_fneg_fabs_lo_fneg_hi:
; GCN: {{buffer|flat}}_load_dword [[A:v[0-9]+]]
; GCN: {{buffer|flat}}_load_dword [[B:v[0-9]+]]
-; GCN: v_cvt_pkrtz_f16_f32_e64 v{{[0-9]+}}, -|[[A]]|, -[[B]]
+; GCN: v_cvt_pkrtz_f16_f32{{(_e64)*}} v{{[0-9]+}}, -|[[A]]|, -[[B]]
define amdgpu_kernel void @v_cvt_pkrtz_v2f16_f32_fneg_fabs_lo_fneg_hi(<2 x half> addrspace(1)* %out, float addrspace(1)* %a.ptr, float addrspace(1)* %b.ptr) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%tid.ext = sext i32 %tid to i64
diff --git a/test/CodeGen/AMDGPU/llvm.amdgcn.mbcnt.ll b/test/CodeGen/AMDGPU/llvm.amdgcn.mbcnt.ll
index ab76c870796be..144c8f428ab0f 100644
--- a/test/CodeGen/AMDGPU/llvm.amdgcn.mbcnt.ll
+++ b/test/CodeGen/AMDGPU/llvm.amdgcn.mbcnt.ll
@@ -2,9 +2,9 @@
; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
; GCN-LABEL: {{^}}mbcnt_intrinsics:
-; GCN: v_mbcnt_lo_u32_b32_e64 [[LO:v[0-9]+]], -1, 0
+; GCN: v_mbcnt_lo_u32_b32{{(_e64)*}} [[LO:v[0-9]+]], -1, 0
; SI: v_mbcnt_hi_u32_b32_e32 {{v[0-9]+}}, -1, [[LO]]
-; VI: v_mbcnt_hi_u32_b32_e64 {{v[0-9]+}}, -1, [[LO]]
+; VI: v_mbcnt_hi_u32_b32 {{v[0-9]+}}, -1, [[LO]]
define amdgpu_ps void @mbcnt_intrinsics(<16 x i8> addrspace(2)* inreg %arg, <16 x i8> addrspace(2)* inreg %arg1, <32 x i8> addrspace(2)* inreg %arg2, i32 inreg %arg3) {
main_body:
%lo = call i32 @llvm.amdgcn.mbcnt.lo(i32 -1, i32 0) #0
diff --git a/test/CodeGen/AMDGPU/madak.ll b/test/CodeGen/AMDGPU/madak.ll
index eb4066a2a0a80..5f1fb0e2d7324 100644
--- a/test/CodeGen/AMDGPU/madak.ll
+++ b/test/CodeGen/AMDGPU/madak.ll
@@ -9,7 +9,7 @@ declare float @llvm.fabs.f32(float) nounwind readnone
; GCN-LABEL: {{^}}madak_f32:
; GCN: buffer_load_dword [[VA:v[0-9]+]]
; GCN: buffer_load_dword [[VB:v[0-9]+]]
-; GCN: v_madak_f32_e32 {{v[0-9]+}}, [[VA]], [[VB]], 0x41200000
+; GCN: v_madak_f32 {{v[0-9]+}}, [[VA]], [[VB]], 0x41200000
define amdgpu_kernel void @madak_f32(float addrspace(1)* noalias %out, float addrspace(1)* noalias %in.a, float addrspace(1)* noalias %in.b) nounwind {
%tid = tail call i32 @llvm.amdgcn.workitem.id.x() nounwind readnone
%in.a.gep = getelementptr float, float addrspace(1)* %in.a, i32 %tid
@@ -63,7 +63,7 @@ define amdgpu_kernel void @madak_2_use_f32(float addrspace(1)* noalias %out, flo
; GCN-LABEL: {{^}}madak_m_inline_imm_f32:
; GCN: buffer_load_dword [[VA:v[0-9]+]]
-; GCN: v_madak_f32_e32 {{v[0-9]+}}, 4.0, [[VA]], 0x41200000
+; GCN: v_madak_f32 {{v[0-9]+}}, 4.0, [[VA]], 0x41200000
define amdgpu_kernel void @madak_m_inline_imm_f32(float addrspace(1)* noalias %out, float addrspace(1)* noalias %in.a) nounwind {
%tid = tail call i32 @llvm.amdgcn.workitem.id.x() nounwind readnone
%in.a.gep = getelementptr float, float addrspace(1)* %in.a, i32 %tid
@@ -198,7 +198,7 @@ define amdgpu_kernel void @no_madak_src1_modifier_f32(float addrspace(1)* noalia
; GCN: s_load_dword [[SGPR0:s[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, {{0xa|0x28}}
; GCN: v_mov_b32_e32 [[SGPR0_VCOPY:v[0-9]+]], [[SGPR0]]
; GCN: buffer_load_dword [[VGPR:v[0-9]+]]
-; GCN: v_madak_f32_e32 [[MADAK:v[0-9]+]], 0.5, [[SGPR0_VCOPY]], 0x42280000
+; GCN: v_madak_f32 [[MADAK:v[0-9]+]], 0.5, [[SGPR0_VCOPY]], 0x42280000
; GCN: v_mul_f32_e32 [[MUL:v[0-9]+]], [[VGPR]], [[MADAK]]
; GCN: buffer_store_dword [[MUL]]
define amdgpu_kernel void @madak_constant_bus_violation(i32 %arg1, float %sgpr0, float %sgpr1) #0 {
diff --git a/test/CodeGen/AMDGPU/promote-alloca-volatile.ll b/test/CodeGen/AMDGPU/promote-alloca-volatile.ll
index 9c43a6dc60f41..d7655993a2d9f 100644
--- a/test/CodeGen/AMDGPU/promote-alloca-volatile.ll
+++ b/test/CodeGen/AMDGPU/promote-alloca-volatile.ll
@@ -1,26 +1,26 @@
; RUN: opt -S -mtriple=amdgcn-unknown-amdhsa -amdgpu-promote-alloca < %s | FileCheck %s
; CHECK-LABEL: @volatile_load(
-; CHECK: alloca [5 x i32]
+; CHECK: alloca [4 x i32]
; CHECK: load volatile i32, i32*
define amdgpu_kernel void @volatile_load(i32 addrspace(1)* nocapture %out, i32 addrspace(1)* nocapture %in) {
entry:
- %stack = alloca [5 x i32], align 4
+ %stack = alloca [4 x i32], align 4
%tmp = load i32, i32 addrspace(1)* %in, align 4
- %arrayidx1 = getelementptr inbounds [5 x i32], [5 x i32]* %stack, i32 0, i32 %tmp
+ %arrayidx1 = getelementptr inbounds [4 x i32], [4 x i32]* %stack, i32 0, i32 %tmp
%load = load volatile i32, i32* %arrayidx1
store i32 %load, i32 addrspace(1)* %out
ret void
}
; CHECK-LABEL: @volatile_store(
-; CHECK: alloca [5 x i32]
+; CHECK: alloca [4 x i32]
; CHECK: store volatile i32 %tmp, i32*
define amdgpu_kernel void @volatile_store(i32 addrspace(1)* nocapture %out, i32 addrspace(1)* nocapture %in) {
entry:
- %stack = alloca [5 x i32], align 4
+ %stack = alloca [4 x i32], align 4
%tmp = load i32, i32 addrspace(1)* %in, align 4
- %arrayidx1 = getelementptr inbounds [5 x i32], [5 x i32]* %stack, i32 0, i32 %tmp
+ %arrayidx1 = getelementptr inbounds [4 x i32], [4 x i32]* %stack, i32 0, i32 %tmp
store volatile i32 %tmp, i32* %arrayidx1
ret void
}
diff --git a/test/CodeGen/AMDGPU/v_madak_f16.ll b/test/CodeGen/AMDGPU/v_madak_f16.ll
index bfb10503aaea2..0148ff470b783 100644
--- a/test/CodeGen/AMDGPU/v_madak_f16.ll
+++ b/test/CodeGen/AMDGPU/v_madak_f16.ll
@@ -4,7 +4,7 @@
; GCN-LABEL: {{^}}madak_f16
; GCN: buffer_load_ushort v[[A_F16:[0-9]+]]
; GCN: buffer_load_ushort v[[B_F16:[0-9]+]]
-; VI: v_madak_f16_e32 v[[R_F16:[0-9]+]], v[[A_F16]], v[[B_F16]], 0x4900{{$}}
+; VI: v_madak_f16 v[[R_F16:[0-9]+]], v[[A_F16]], v[[B_F16]], 0x4900{{$}}
; VI: buffer_store_short v[[R_F16]]
; GCN: s_endpgm
define amdgpu_kernel void @madak_f16(
diff --git a/test/CodeGen/AMDGPU/waitcnt.mir b/test/CodeGen/AMDGPU/waitcnt.mir
index 38662e83b359d..f754415dccb4b 100644
--- a/test/CodeGen/AMDGPU/waitcnt.mir
+++ b/test/CodeGen/AMDGPU/waitcnt.mir
@@ -51,21 +51,21 @@ name: flat_zero_waitcnt
body: |
bb.0:
successors: %bb.1
- %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, 0, implicit %exec, implicit %flat_scr :: (load 4 from %ir.global4)
- %vgpr3_vgpr4_vgpr5_vgpr6 = FLAT_LOAD_DWORDX4 %vgpr7_vgpr8, 0, 0, 0, implicit %exec, implicit %flat_scr :: (load 16 from %ir.global16)
+ %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, implicit %exec, implicit %flat_scr :: (load 4 from %ir.global4)
+ %vgpr3_vgpr4_vgpr5_vgpr6 = FLAT_LOAD_DWORDX4 %vgpr7_vgpr8, 0, 0, implicit %exec, implicit %flat_scr :: (load 16 from %ir.global16)
%vgpr0 = V_MOV_B32_e32 %vgpr1, implicit %exec
S_BRANCH %bb.1
bb.1:
successors: %bb.2
- %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, 0, implicit %exec, implicit %flat_scr
- %vgpr3_vgpr4_vgpr5_vgpr6 = FLAT_LOAD_DWORDX4 %vgpr7_vgpr8, 0, 0, 0, implicit %exec, implicit %flat_scr :: (load 16 from %ir.global16)
+ %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, implicit %exec, implicit %flat_scr
+ %vgpr3_vgpr4_vgpr5_vgpr6 = FLAT_LOAD_DWORDX4 %vgpr7_vgpr8, 0, 0, implicit %exec, implicit %flat_scr :: (load 16 from %ir.global16)
%vgpr0 = V_MOV_B32_e32 %vgpr1, implicit %exec
S_BRANCH %bb.2
bb.2:
- %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, 0, implicit %exec, implicit %flat_scr :: (load 4 from %ir.flat4)
- %vgpr3_vgpr4_vgpr5_vgpr6 = FLAT_LOAD_DWORDX4 %vgpr7_vgpr8, 0, 0, 0, implicit %exec, implicit %flat_scr :: (load 16 from %ir.flat16)
+ %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, implicit %exec, implicit %flat_scr :: (load 4 from %ir.flat4)
+ %vgpr3_vgpr4_vgpr5_vgpr6 = FLAT_LOAD_DWORDX4 %vgpr7_vgpr8, 0, 0, implicit %exec, implicit %flat_scr :: (load 16 from %ir.flat16)
%vgpr0 = V_MOV_B32_e32 %vgpr1, implicit %exec
S_ENDPGM
...
@@ -86,11 +86,11 @@ name: single_fallthrough_successor_no_end_block_wait
body: |
bb.0:
successors: %bb.1
- %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, 0, implicit %exec, implicit %flat_scr
+ %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, implicit %exec, implicit %flat_scr
bb.1:
%vgpr3_vgpr4 = V_LSHLREV_B64 4, %vgpr7_vgpr8, implicit %exec
- FLAT_STORE_DWORD %vgpr3_vgpr4, %vgpr0, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORD %vgpr3_vgpr4, %vgpr0, 0, 0, implicit %exec, implicit %flat_scr
S_ENDPGM
...
---
@@ -114,15 +114,15 @@ name: single_branch_successor_not_next_block
body: |
bb.0:
successors: %bb.2
- %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, 0, implicit %exec, implicit %flat_scr
+ %vgpr0 = FLAT_LOAD_DWORD %vgpr1_vgpr2, 0, 0, implicit %exec, implicit %flat_scr
S_BRANCH %bb.2
bb.1:
- FLAT_STORE_DWORD %vgpr8_vgpr9, %vgpr10, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORD %vgpr8_vgpr9, %vgpr10, 0, 0, implicit %exec, implicit %flat_scr
S_ENDPGM
bb.2:
%vgpr3_vgpr4 = V_LSHLREV_B64 4, %vgpr7_vgpr8, implicit %exec
- FLAT_STORE_DWORD %vgpr3_vgpr4, %vgpr0, 0, 0, 0, implicit %exec, implicit %flat_scr
+ FLAT_STORE_DWORD %vgpr3_vgpr4, %vgpr0, 0, 0, implicit %exec, implicit %flat_scr
S_ENDPGM
...
generated by cgit v1.3 (git 2.53.0) at 2026-09-07 19:04:07 +0000