File: shifts.ll

package info (click to toggle)
llvm-toolchain-9 1%3A9.0.1-16.1
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 882,388 kB
  • sloc: cpp: 4,167,636; ansic: 714,256; asm: 457,610; python: 155,927; objc: 65,094; sh: 42,856; lisp: 26,908; perl: 7,786; pascal: 7,722; makefile: 6,881; ml: 5,581; awk: 3,648; cs: 2,027; xml: 888; javascript: 381; ruby: 156
file content (61 lines) | stat: -rw-r--r-- 2,202 bytes parent folder | download | duplicates (6)
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
61
; RUN: opt -cost-model -analyze -mtriple=amdgcn-unknown-amdhsa -mattr=+half-rate-64-ops < %s | FileCheck -check-prefix=ALL -check-prefix=FAST64 %s
; RUN: opt -cost-model -analyze -mtriple=amdgcn-unknown-amdhsa -mattr=-half-rate-64-ops < %s | FileCheck -check-prefix=ALL -check-prefix=SLOW64 %s

; ALL: 'shl_i32'
; ALL: estimated cost of 1 for {{.*}} shl i32
define amdgpu_kernel void @shl_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %vaddr, i32 %b) #0 {
  %vec = load i32, i32 addrspace(1)* %vaddr
  %or = shl i32 %vec, %b
  store i32 %or, i32 addrspace(1)* %out
  ret void
}

; ALL: 'shl_i64'
; FAST64: estimated cost of 2 for {{.*}} shl i64
; SLOW64: estimated cost of 3 for {{.*}} shl i64
define amdgpu_kernel void @shl_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %vaddr, i64 %b) #0 {
  %vec = load i64, i64 addrspace(1)* %vaddr
  %or = shl i64 %vec, %b
  store i64 %or, i64 addrspace(1)* %out
  ret void
}

; ALL: 'lshr_i32'
; ALL: estimated cost of 1 for {{.*}} lshr i32
define amdgpu_kernel void @lshr_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %vaddr, i32 %b) #0 {
  %vec = load i32, i32 addrspace(1)* %vaddr
  %or = lshr i32 %vec, %b
  store i32 %or, i32 addrspace(1)* %out
  ret void
}

; ALL: 'lshr_i64'
; FAST64: estimated cost of 2 for {{.*}} lshr i64
; SLOW64: estimated cost of 3 for {{.*}} lshr i64
define amdgpu_kernel void @lshr_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %vaddr, i64 %b) #0 {
  %vec = load i64, i64 addrspace(1)* %vaddr
  %or = lshr i64 %vec, %b
  store i64 %or, i64 addrspace(1)* %out
  ret void
}

; ALL: 'ashr_i32'
; ALL: estimated cost of 1 for {{.*}} ashr i32
define amdgpu_kernel void @ashr_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %vaddr, i32 %b) #0 {
  %vec = load i32, i32 addrspace(1)* %vaddr
  %or = ashr i32 %vec, %b
  store i32 %or, i32 addrspace(1)* %out
  ret void
}

; ALL: 'ashr_i64'
; FAST64: estimated cost of 2 for {{.*}} ashr i64
; SLOW64: estimated cost of 3 for {{.*}} ashr i64
define amdgpu_kernel void @ashr_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %vaddr, i64 %b) #0 {
  %vec = load i64, i64 addrspace(1)* %vaddr
  %or = ashr i64 %vec, %b
  store i64 %or, i64 addrspace(1)* %out
  ret void
}

attributes #0 = { nounwind }