Emit llvm.loop.parallel_accesses metadata instead of llvm.mem.parallel_loop_access. The latter is deprecated because it assumes that LoopIDs are persistent, which they are not. We also emit parallel access metadata for all surrounding parallel loops, not just the innermost parallel.
40 lines
1.5 KiB
LLVM
40 lines
1.5 KiB
LLVM
; RUN: opt %loadPolly -polly-vectorizer=polly -polly-codegen \
|
|
; RUN: < %s -S | FileCheck %s
|
|
|
|
; #pragma known-parallel
|
|
; for (int c0 = 0; c0 <= min(15, N - 1); c0 += 1)
|
|
; Stmt_if_then(c0);
|
|
|
|
; CHECK: polly.stmt.if.then: ; preds = %polly.loop_header
|
|
; CHECK: %p_conv = sitofp i64 %polly.indvar to float
|
|
; CHECK: %scevgep = getelementptr float, float* %A, i64 %polly.indvar
|
|
; CHECK: %_p_scalar_ = load float, float* %scevgep, align 4, !alias.scope !0, !noalias !2, !llvm.access.group !3
|
|
; CHECK: %p_add = fadd float %p_conv, %_p_scalar_
|
|
; CHECK: store float %p_add, float* %scevgep, align 4, !alias.scope !0, !noalias !2, !llvm.access.group !3
|
|
|
|
define void @foo(float* %A, i64 %N) #0 {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body: ; preds = %entry, %for.inc
|
|
%i.02 = phi i64 [ 0, %entry ], [ %inc, %for.inc ]
|
|
%cmp1 = icmp slt i64 %i.02, %N
|
|
br i1 %cmp1, label %if.then, label %for.inc
|
|
|
|
if.then: ; preds = %for.body
|
|
%conv = sitofp i64 %i.02 to float
|
|
%arrayidx = getelementptr inbounds float, float* %A, i64 %i.02
|
|
%0 = load float, float* %arrayidx, align 4
|
|
%add = fadd float %conv, %0
|
|
store float %add, float* %arrayidx, align 4
|
|
br label %for.inc
|
|
|
|
for.inc: ; preds = %for.body, %if.then
|
|
%inc = add nuw nsw i64 %i.02, 1
|
|
%exitcond = icmp ne i64 %inc, 16
|
|
br i1 %exitcond, label %for.body, label %for.end
|
|
|
|
for.end: ; preds = %for.inc
|
|
ret void
|
|
}
|