1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-22 04:22:57 +02:00
llvm-mirror/test/CodeGen/AMDGPU/branch-uniformity.ll
Matt Arsenault 3a9a1ac61b AMDGPU: Use unsigned compare for eq/ne
For some reason there are both of these available, except
for scalar 64-bit compares which only has u64. I'm not sure
why there are both (I'm guessing it's for the one bit inputs we
don't use), but for consistency always using the
unsigned one.

llvm-svn: 282832
2016-09-30 01:50:20 +00:00

42 lines
1.5 KiB
LLVM

; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck %s
; The branch instruction in LOOP49 has a uniform condition, but PHI instructions
; introduced by the structurizecfg pass previously caused a false divergence
; which ended up in an assertion (or incorrect code) because
; SIAnnotateControlFlow and structurizecfg had different ideas about which
; branches are uniform.
;
; CHECK-LABEL: {{^}}main:
; CHECK: ; %LOOP49
; CHECK: v_cmp_ne_u32_e32 vcc,
; CHECK: s_cbranch_vccnz
; CHECK: ; %ENDIF53
define amdgpu_vs float @main(i32 %in) {
main_body:
%cmp = mul i32 %in, 2
br label %LOOP
LOOP: ; preds = %ENDLOOP48, %main_body
%counter = phi i32 [ 0, %main_body ], [ %counter.next, %ENDLOOP48 ]
%v.LOOP = phi i32 [ 0, %main_body ], [ %v.ENDLOOP48, %ENDLOOP48 ]
%tmp7 = icmp slt i32 %cmp, %counter
br i1 %tmp7, label %IF, label %LOOP49
IF: ; preds = %LOOP
%r = bitcast i32 %v.LOOP to float
ret float %r
LOOP49: ; preds = %LOOP
%tmp8 = icmp ne i32 %counter, 0
br i1 %tmp8, label %ENDLOOP48, label %ENDIF53
ENDLOOP48: ; preds = %ENDIF53, %LOOP49
%v.ENDLOOP48 = phi i32 [ %v.LOOP, %LOOP49 ], [ %v.ENDIF53, %ENDIF53 ]
%counter.next = add i32 %counter, 1
br label %LOOP
ENDIF53: ; preds = %LOOP49
%v.ENDIF53 = add i32 %v.LOOP, %counter
br label %ENDLOOP48
}