File: v_cvt_pk_u8_f32.ll

package info (click to toggle)
llvm-toolchain-19 1%3A19.1.7-3~deb12u1
  • links: PTS, VCS
  • area: main
  • in suites: bookworm-proposed-updates
  • size: 1,998,492 kB
  • sloc: cpp: 6,951,680; ansic: 1,486,157; asm: 913,598; python: 232,024; f90: 80,126; objc: 75,281; lisp: 37,276; pascal: 16,990; sh: 10,009; ml: 5,058; perl: 4,724; awk: 3,523; makefile: 3,167; javascript: 2,504; xml: 892; fortran: 664; cs: 573
file content (60 lines) | stat: -rw-r--r-- 2,838 bytes parent folder | download | duplicates (7)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
; RUN: llc -mtriple=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
; RUN: llc -mtriple=amdgcn -mcpu=fiji -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s

declare i32 @llvm.amdgcn.cvt.pk.u8.f32(float, i32, i32) #0

; GCN-LABEL: {{^}}v_cvt_pk_u8_f32_idx_0:
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 0, v{{[0-9]+}}
define amdgpu_kernel void @v_cvt_pk_u8_f32_idx_0(ptr addrspace(1) %out, float %src, i32 %reg) {
  %result = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 0, i32 %reg) #0
  store i32 %result, ptr addrspace(1) %out, align 4
  ret void
}

; GCN-LABEL: {{^}}v_cvt_pk_u8_f32_idx_1:
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 1, v{{[0-9]+}}
define amdgpu_kernel void @v_cvt_pk_u8_f32_idx_1(ptr addrspace(1) %out, float %src, i32 %reg) {
  %result = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 1, i32 %reg) #0
  store i32 %result, ptr addrspace(1) %out, align 4
  ret void
}

; GCN-LABEL: {{^}}v_cvt_pk_u8_f32_idx_2:
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 2, v{{[0-9]+}}
define amdgpu_kernel void @v_cvt_pk_u8_f32_idx_2(ptr addrspace(1) %out, float %src, i32 %reg) {
  %result = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 2, i32 %reg) #0
  store i32 %result, ptr addrspace(1) %out, align 4
  ret void
}

; GCN-LABEL: {{^}}v_cvt_pk_u8_f32_idx_3:
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 3, v{{[0-9]+}}
define amdgpu_kernel void @v_cvt_pk_u8_f32_idx_3(ptr addrspace(1) %out, float %src, i32 %reg) {
  %result = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 3, i32 %reg) #0
  store i32 %result, ptr addrspace(1) %out, align 4
  ret void
}

; GCN-LABEL: {{^}}v_cvt_pk_u8_f32_combine:
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 0, v{{[0-9]+}}
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 1, v{{[0-9]+}}
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 2, v{{[0-9]+}}
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, 3, v{{[0-9]+}}
define amdgpu_kernel void @v_cvt_pk_u8_f32_combine(ptr addrspace(1) %out, float %src, i32 %reg) {
  %result0 = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 0, i32 %reg) #0
  %result1 = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 1, i32 %result0) #0
  %result2 = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 2, i32 %result1) #0
  %result3 = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 3, i32 %result2) #0
  store i32 %result3, ptr addrspace(1) %out, align 4
  ret void
}

; GCN-LABEL: {{^}}v_cvt_pk_u8_f32_idx:
; GCN: v_cvt_pk_u8_f32 v{{[0-9]+}}, s{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}
define amdgpu_kernel void @v_cvt_pk_u8_f32_idx(ptr addrspace(1) %out, float %src, i32 %idx, i32 %reg) {
  %result = call i32 @llvm.amdgcn.cvt.pk.u8.f32(float %src, i32 %idx, i32 %reg) #0
  store i32 %result, ptr addrspace(1) %out, align 4
  ret void
}

attributes #0 = { nounwind readnone }