File: divergence-at-use.ll

package info (click to toggle)
llvm-toolchain-15 1%3A15.0.6-4
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,554,644 kB
  • sloc: cpp: 5,922,452; ansic: 1,012,136; asm: 674,362; python: 191,568; objc: 73,855; f90: 42,327; lisp: 31,913; pascal: 11,973; javascript: 10,144; sh: 9,421; perl: 7,447; ml: 5,527; awk: 3,523; makefile: 2,520; xml: 885; cs: 573; fortran: 567
file content (20 lines) | stat: -rw-r--r-- 653 bytes parent folder | download | duplicates (9) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
 
; RUN: llc -march=amdgcn -mcpu=gfx900 -amdgpu-atomic-optimizations=true < %s | FileCheck %s
; RUN: llc -march=amdgcn -mcpu=gfx900 -amdgpu-atomic-optimizations=true < %s -use-gpu-divergence-analysis | FileCheck %s

@local = addrspace(3) global i32 undef

define amdgpu_kernel void @reducible(i32 %x) {
; CHECK-LABEL: reducible:
; CHECK-NOT: dpp
entry:
  br label %loop
loop:
  %i = phi i32 [ 0, %entry ], [ %i1, %loop ]
  %gep = getelementptr i32, i32 addrspace(3)* @local, i32 %i
  %cond = icmp ult i32 %i, %x
  %i1 = add i32 %i, 1
  br i1 %cond, label %loop, label %exit
exit:
  %old = atomicrmw add i32 addrspace(3)* %gep, i32 %x acq_rel
  ret void
}