[NFC][AMDGPU] Move image-atomic-attributes test to test/Assembler.

[sstipano]

No description provided.

[llvmbot]

@llvm/pr-subscribers-backend-amdgpu

Author: None (sstipanovic)

Changes

---

Patch is 27.29 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/76917.diff

2 Files Affected:

• (added) llvm/test/Assembler/amdgpu-image-atomic-attributes.ll (+195)
• (removed) llvm/test/CodeGen/AMDGPU/image-atomic-attributes.ll (-356)

diff --git a/llvm/test/Assembler/amdgpu-image-atomic-attributes.ll b/llvm/test/Assembler/amdgpu-image-atomic-attributes.ll
new file mode 100644
index 00000000000000..3dc32e40308196
--- /dev/null
+++ b/llvm/test/Assembler/amdgpu-image-atomic-attributes.ll
@@ -0,0 +1,195 @@
+; RUN: opt -S -mtriple=amdgcn-unknown-unknown < %s | FileCheck -check-prefixes=CHECK %s
+
+define amdgpu_ps float @atomic_swap_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.swap.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps <2 x float> @atomic_swap_1d_i64(<8 x i32> inreg %rsrc, i64 %data, i32 %s) {
+main_body:
+  %v = call i64 @llvm.amdgcn.image.atomic.swap.1d.i64.i32(i64 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i64 %v to <2 x float>
+  ret <2 x float> %out
+}
+
+define amdgpu_ps float @atomic_add_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_sub_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.sub.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_smin_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.smin.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_umin_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.umin.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_smax_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.smax.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_umax_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.umax.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_and_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.and.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_or_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.or.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_xor_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.xor.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_inc_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.inc.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_dec_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.dec.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_cmpswap_1d(<8 x i32> inreg %rsrc, i32 %cmp, i32 %swap, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.cmpswap.1d.i32.i32(i32 %cmp, i32 %swap, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps <2 x float> @atomic_cmpswap_1d_64(<8 x i32> inreg %rsrc, i64 %cmp, i64 %swap, i32 %s) {
+main_body:
+  %v = call i64 @llvm.amdgcn.image.atomic.cmpswap.1d.i64.i32(i64 %cmp, i64 %swap, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i64 %v to <2 x float>
+  ret <2 x float> %out
+}
+
+define amdgpu_ps float @atomic_add_2d(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.2d.i32.i32(i32 %data, i32 %s, i32 %t, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_add_3d(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t, i32 %r) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.3d.i32.i32(i32 %data, i32 %s, i32 %t, i32 %r, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_add_cube(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t, i32 %face) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.cube.i32.i32(i32 %data, i32 %s, i32 %t, i32 %face, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_add_1darray(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %slice) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.1darray.i32.i32(i32 %data, i32 %s, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_add_2darray(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t, i32 %slice) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.2darray.i32.i32(i32 %data, i32 %s, i32 %t, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_add_2dmsaa(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t, i32 %fragid) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.2dmsaa.i32.i32(i32 %data, i32 %s, i32 %t, i32 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_add_2darraymsaa(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t, i32 %slice, i32 %fragid) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.2darraymsaa.i32.i32(i32 %data, i32 %s, i32 %t, i32 %slice, i32 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+define amdgpu_ps float @atomic_add_1d_slc(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
+main_body:
+  %v = call i32 @llvm.amdgcn.image.atomic.add.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 2)
+  %out = bitcast i32 %v to float
+  ret float %out
+}
+
+declare i32 @llvm.amdgcn.image.atomic.swap.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.add.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.sub.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.smin.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.umin.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.smax.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.umax.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.and.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.or.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.xor.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.inc.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.dec.1d.i32.i32(i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.cmpswap.1d.i32.i32(i32, i32, i32, <8 x i32>, i32, i32) #0
+
+declare i64 @llvm.amdgcn.image.atomic.swap.1d.i64.i32(i64, i32, <8 x i32>, i32, i32) #0
+declare i64 @llvm.amdgcn.image.atomic.cmpswap.1d.i64.i32(i64, i64, i32, <8 x i32>, i32, i32) #0
+
+declare i32 @llvm.amdgcn.image.atomic.add.2d.i32.i32(i32, i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.add.3d.i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.add.cube.i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.add.1darray.i32.i32(i32, i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.add.2darray.i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.add.2dmsaa.i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #0
+declare i32 @llvm.amdgcn.image.atomic.add.2darraymsaa.i32.i32(i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
+
+attributes #0 = { nounwind }
+attributes #1 = { nounwind readonly }
+attributes #2 = { nounwind readnone }
+
+;.
+; CHECK: attributes #[[ATTR0:[0-9]+]] = { nocallback nofree nounwind willreturn }
+;.
\ No newline at end of file
diff --git a/llvm/test/CodeGen/AMDGPU/image-atomic-attributes.ll b/llvm/test/CodeGen/AMDGPU/image-atomic-attributes.ll
deleted file mode 100644
index dd60d9d702716d..00000000000000
--- a/llvm/test/CodeGen/AMDGPU/image-atomic-attributes.ll
+++ /dev/null
@@ -1,356 +0,0 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-attributes --check-globals all --version 4
-; RUN: opt -S -mtriple=amdgcn-unknown-unknown < %s | FileCheck -check-prefixes=CHECK %s
-
-define amdgpu_ps float @atomic_swap_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_swap_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.swap.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.swap.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps <2 x float> @atomic_swap_1d_i64(<8 x i32> inreg %rsrc, i64 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps <2 x float> @atomic_swap_1d_i64(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i64 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i64 @llvm.amdgcn.image.atomic.swap.1d.i64.i32(i64 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i64 [[V]] to <2 x float>
-; CHECK-NEXT:    ret <2 x float> [[OUT]]
-;
-main_body:
-  %v = call i64 @llvm.amdgcn.image.atomic.swap.1d.i64.i32(i64 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i64 %v to <2 x float>
-  ret <2 x float> %out
-}
-
-define amdgpu_ps float @atomic_add_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_add_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.add.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.add.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_sub_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_sub_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.sub.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.sub.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_smin_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_smin_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.smin.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.smin.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_umin_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_umin_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.umin.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.umin.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_smax_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_smax_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.smax.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.smax.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_umax_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_umax_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.umax.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.umax.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_and_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_and_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.and.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.and.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_or_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_or_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.or.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.or.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_xor_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_xor_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.xor.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.xor.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_inc_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_inc_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.inc.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.inc.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_dec_1d(<8 x i32> inreg %rsrc, i32 %data, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_dec_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.dec.1d.i32.i32(i32 [[DATA]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.dec.1d.i32.i32(i32 %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_cmpswap_1d(<8 x i32> inreg %rsrc, i32 %cmp, i32 %swap, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_cmpswap_1d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[CMP:%.*]], i32 [[SWAP:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.cmpswap.1d.i32.i32(i32 [[CMP]], i32 [[SWAP]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.cmpswap.1d.i32.i32(i32 %cmp, i32 %swap, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps <2 x float> @atomic_cmpswap_1d_64(<8 x i32> inreg %rsrc, i64 %cmp, i64 %swap, i32 %s) {
-; CHECK-LABEL: define amdgpu_ps <2 x float> @atomic_cmpswap_1d_64(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i64 [[CMP:%.*]], i64 [[SWAP:%.*]], i32 [[S:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i64 @llvm.amdgcn.image.atomic.cmpswap.1d.i64.i32(i64 [[CMP]], i64 [[SWAP]], i32 [[S]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i64 [[V]] to <2 x float>
-; CHECK-NEXT:    ret <2 x float> [[OUT]]
-;
-main_body:
-  %v = call i64 @llvm.amdgcn.image.atomic.cmpswap.1d.i64.i32(i64 %cmp, i64 %swap, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i64 %v to <2 x float>
-  ret <2 x float> %out
-}
-
-define amdgpu_ps float @atomic_add_2d(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t) {
-; CHECK-LABEL: define amdgpu_ps float @atomic_add_2d(
-; CHECK-SAME: <8 x i32> inreg [[RSRC:%.*]], i32 [[DATA:%.*]], i32 [[S:%.*]], i32 [[T:%.*]]) {
-; CHECK-NEXT:  main_body:
-; CHECK-NEXT:    [[V:%.*]] = call i32 @llvm.amdgcn.image.atomic.add.2d.i32.i32(i32 [[DATA]], i32 [[S]], i32 [[T]], <8 x i32> [[RSRC]], i32 0, i32 0)
-; CHECK-NEXT:    [[OUT:%.*]] = bitcast i32 [[V]] to float
-; CHECK-NEXT:    ret float [[OUT]]
-;
-main_body:
-  %v = call i32 @llvm.amdgcn.image.atomic.add.2d.i32.i32(i32 %data, i32 %s, i32 %t, <8 x i32> %rsrc, i32 0, i32 0)
-  %out = bitcast i32 %v to float
-  ret float %out
-}
-
-define amdgpu_ps float @atomic_add_3d(<8 x i32> inreg %rsrc, i32 %data, i32 %s, i32 %t, i32 %r) {
-; CHECK-L...
[truncated]

[arsenm]

Drop the pre-declared wrong attributes

[arsenm]

-check-prefixes=CHECK is the same as the default, can just drop it