; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn -mcpu=gfx90a -amdgpu-atomic-optimizer-strategy=None -o - %s | FileCheck %s

%S = type <{ float, double }>

; The result of that atomic ops should not be used as a uniform value.

define protected amdgpu_kernel void @add(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: add:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_add v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw add ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @sub(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: sub:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_sub v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw sub ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @and(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: and:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_and v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw and ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @or(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: or:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_or v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw or ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @xor(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: xor:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_xor v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw xor ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @nand(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: nand:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    s_mov_b64 s[4:5], 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    s_load_dword s6, s[0:1], 0x0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v0, s6
; CHECK-NEXT:  .LBB5_1: ; %atomicrmw.start
; CHECK-NEXT:    ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    v_mov_b32_e32 v3, v0
; CHECK-NEXT:    v_not_b32_e32 v0, v3
; CHECK-NEXT:    v_or_b32_e32 v2, -2, v0
; CHECK-NEXT:    global_atomic_cmpswap v0, v1, v[2:3], s[0:1] glc
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cmp_eq_u32_e32 vcc, v0, v3
; CHECK-NEXT:    s_or_b64 s[4:5], vcc, s[4:5]
; CHECK-NEXT:    s_andn2_b64 exec, exec, s[4:5]
; CHECK-NEXT:    s_cbranch_execnz .LBB5_1
; CHECK-NEXT:  ; %bb.2: ; %atomicrmw.end
; CHECK-NEXT:    s_or_b64 exec, exec, s[4:5]
; CHECK-NEXT:    v_mov_b32_e32 v2, s2
; CHECK-NEXT:    v_mov_b32_e32 v3, s3
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, v[2:3]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw nand ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @max_workgroup(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: max_workgroup:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_smax v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw max ptr addrspace(1) %p, i32 1 syncscope("workgroup") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @max(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: max:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_smax v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw max ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @min_workgroup(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: min_workgroup:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_smin v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw min ptr addrspace(1) %p, i32 1 syncscope("workgroup") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @min(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: min:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_smin v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw min ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @umax_workgroup(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: umax_workgroup:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_umax v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw umax ptr addrspace(1) %p, i32 1 syncscope("workgroup") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @umax(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: umax:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_umax v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw umax ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @umin_workgroup(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: umin_workgroup:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_umin v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw umin ptr addrspace(1) %p, i32 1 syncscope("workgroup") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @umin(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: umin:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_umin v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw umin ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @cmpxchg(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: cmpxchg:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v2, 0
; CHECK-NEXT:    v_mov_b32_e32 v0, 2
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_cmpswap v2, v2, v[0:1], s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %agg = cmpxchg ptr addrspace(1) %p, i32 1, i32 2 monotonic monotonic
  %n32 = extractvalue {i32, i1} %agg, 0
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @xchg(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: xchg:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_swap v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw xchg ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @inc(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: inc:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_inc v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw uinc_wrap ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @dec(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: dec:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_dec v2, v0, v1, s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = atomicrmw udec_wrap ptr addrspace(1) %p, i32 1 syncscope("agent") monotonic
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @fadd(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: fadd:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    s_mov_b64 s[4:5], 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    s_load_dword s6, s[0:1], 0x0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v0, s6
; CHECK-NEXT:  .LBB18_1: ; %atomicrmw.start
; CHECK-NEXT:    ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    v_mov_b32_e32 v3, v0
; CHECK-NEXT:    v_add_f32_e32 v2, 1.0, v3
; CHECK-NEXT:    global_atomic_cmpswap v0, v1, v[2:3], s[0:1] glc
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cmp_eq_u32_e32 vcc, v0, v3
; CHECK-NEXT:    s_or_b64 s[4:5], vcc, s[4:5]
; CHECK-NEXT:    s_andn2_b64 exec, exec, s[4:5]
; CHECK-NEXT:    s_cbranch_execnz .LBB18_1
; CHECK-NEXT:  ; %bb.2: ; %atomicrmw.end
; CHECK-NEXT:    s_or_b64 exec, exec, s[4:5]
; CHECK-NEXT:    v_cvt_u32_f32_e32 v2, v0
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %f32 = atomicrmw fadd ptr addrspace(1) %p, float 1.0 syncscope("agent") monotonic
  %n32 = fptoui float %f32 to i32
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @fsub(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: fsub:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    s_mov_b64 s[4:5], 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    s_load_dword s6, s[0:1], 0x0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v0, s6
; CHECK-NEXT:  .LBB19_1: ; %atomicrmw.start
; CHECK-NEXT:    ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    v_mov_b32_e32 v3, v0
; CHECK-NEXT:    v_add_f32_e32 v2, -1.0, v3
; CHECK-NEXT:    global_atomic_cmpswap v0, v1, v[2:3], s[0:1] glc
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cmp_eq_u32_e32 vcc, v0, v3
; CHECK-NEXT:    s_or_b64 s[4:5], vcc, s[4:5]
; CHECK-NEXT:    s_andn2_b64 exec, exec, s[4:5]
; CHECK-NEXT:    s_cbranch_execnz .LBB19_1
; CHECK-NEXT:  ; %bb.2: ; %atomicrmw.end
; CHECK-NEXT:    s_or_b64 exec, exec, s[4:5]
; CHECK-NEXT:    v_cvt_u32_f32_e32 v2, v0
; CHECK-NEXT:    v_mov_b32_e32 v0, s2
; CHECK-NEXT:    v_mov_b32_e32 v1, s3
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v2, 12, v[0:1]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %f32 = atomicrmw fsub ptr addrspace(1) %p, float 1.0 syncscope("agent") monotonic
  %n32 = fptoui float %f32 to i32
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @fmin(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: fmin:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v2, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 0x3ff00000
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_min_f64 v[0:1], v2, v[0:1], s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v2, s2
; CHECK-NEXT:    v_mov_b32_e32 v3, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cvt_u32_f64_e32 v0, v[0:1]
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, v[2:3]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm

  %f64 = atomicrmw fmin ptr addrspace(1) %p, double 1.0 syncscope("agent") monotonic, !amdgpu.no.fine.grained.memory !0
  %n32 = fptoui double %f64 to i32
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @fmax(ptr addrspace(1) %p, ptr addrspace(1) %q) {
; CHECK-LABEL: fmax:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v2, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 0x3ff00000
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    global_atomic_max_f64 v[0:1], v2, v[0:1], s[0:1] glc
; CHECK-NEXT:    v_mov_b32_e32 v2, s2
; CHECK-NEXT:    v_mov_b32_e32 v3, s3
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cvt_u32_f64_e32 v0, v[0:1]
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, v[2:3]
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %f64 = atomicrmw fmax ptr addrspace(1) %p, double 1.0 syncscope("agent") monotonic, !amdgpu.no.fine.grained.memory !0
  %n32 = fptoui double %f64 to i32
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.swap(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.swap:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_swap v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.swap.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.add(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.add:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_add v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.add.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.sub(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.sub:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_sub v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.sub.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.smin(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.smin:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_smin v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.smin.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.smax(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.smax:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_smax v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.smax.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.umin(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.umin:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_umin v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.umin.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.umax(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.umax:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_umax v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.umax.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.and(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.and:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_and v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.and.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.or(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.or:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_or v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.or.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.xor(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.xor:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_xor v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.xor.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.inc(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.inc:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_inc v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.inc.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.dec(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.dec:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v1, s6
; CHECK-NEXT:    buffer_atomic_dec v0, v1, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.dec.i32(i32 1, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.cmpswap(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.cmpswap:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v1, 2
; CHECK-NEXT:    v_mov_b32_e32 v0, 1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v2, s6
; CHECK-NEXT:    buffer_atomic_cmpswap v[0:1], v2, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %n32 = call i32 @llvm.amdgcn.raw.ptr.buffer.atomic.cmpswap.i32(i32 1, i32 2, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.fadd(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.fadd:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v1, 1.0
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v0, s6
; CHECK-NEXT:    buffer_atomic_add_f32 v1, v0, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cvt_u32_f32_e32 v0, v1
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %f32 = call float @llvm.amdgcn.raw.ptr.buffer.atomic.fadd.f32(float 1.0, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n32 = fptoui float %f32 to i32
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.fmin(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.fmin:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 0x3ff00000
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v2, s6
; CHECK-NEXT:    buffer_atomic_min_f64 v[0:1], v2, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cvt_u32_f64_e32 v0, v[0:1]
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %f64 = call double @llvm.amdgcn.raw.ptr.buffer.atomic.fmin.f64(double 1.0, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n32 = fptoui double %f64 to i32
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

define protected amdgpu_kernel void @buffer.ptr.atomic.fmax(ptr addrspace(8) %rsrc, i32 %vindex, ptr addrspace(1) %q) {
; CHECK-LABEL: buffer.ptr.atomic.fmax:
; CHECK:       ; %bb.0:
; CHECK-NEXT:    s_load_dword s6, s[4:5], 0x34
; CHECK-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; CHECK-NEXT:    v_mov_b32_e32 v0, 0
; CHECK-NEXT:    v_mov_b32_e32 v1, 0x3ff00000
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mov_b32_e32 v2, s6
; CHECK-NEXT:    buffer_atomic_max_f64 v[0:1], v2, s[0:3], 0 offen glc
; CHECK-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x3c
; CHECK-NEXT:    v_mov_b32_e32 v2, 1.0
; CHECK-NEXT:    s_waitcnt vmcnt(0)
; CHECK-NEXT:    v_cvt_u32_f64_e32 v0, v[0:1]
; CHECK-NEXT:    s_waitcnt lgkmcnt(0)
; CHECK-NEXT:    v_mad_u64_u32 v[0:1], s[0:1], v0, 12, s[0:1]
; CHECK-NEXT:    global_store_dword v[0:1], v2, off
; CHECK-NEXT:    s_endpgm
  %f64 = call double @llvm.amdgcn.raw.ptr.buffer.atomic.fmax.f64(double 1.0, ptr addrspace(8) %rsrc, i32 %vindex, i32 0, i32 0)
  %n32 = fptoui double %f64 to i32
  %n64 = zext i32 %n32 to i64
  %p1 = getelementptr inbounds %S, ptr addrspace(1) %q, i64 %n64, i32 0
  store float 1.0, ptr addrspace(1) %p1
  ret void
}

declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.swap.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.add.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.sub.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.smin.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.smax.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.umin.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.umax.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.and.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.or.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.xor.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.inc.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.dec.i32(i32, ptr addrspace(8), i32, i32, i32)
declare i32 @llvm.amdgcn.raw.ptr.buffer.atomic.cmpswap.i32(i32, i32, ptr addrspace(8), i32, i32, i32)
declare float @llvm.amdgcn.raw.ptr.buffer.atomic.fadd.f32(float, ptr addrspace(8), i32, i32, i32)
declare double @llvm.amdgcn.raw.ptr.buffer.atomic.fmin.f64(double, ptr addrspace(8), i32, i32, i32)
declare double @llvm.amdgcn.raw.ptr.buffer.atomic.fmax.f64(double, ptr addrspace(8), i32, i32, i32)

!0 = !{}