1 ; RUN: opt < %s -S -loop-unroll -mtriple aarch64 -mcpu=cortex-a57 | FileCheck %s -check-prefix=EPILOG 2 ; RUN: opt < %s -S -loop-unroll -mtriple aarch64 -mcpu=cortex-a57 -unroll-runtime-epilog=false | FileCheck %s -check-prefix=PROLOG 3 4 ; Tests for unrolling loops with run-time trip counts 5 6 ; EPILOG: %xtraiter = and i32 %n 7 ; EPILOG: for.body: 8 ; EPILOG: %lcmp.mod = icmp ne i32 %xtraiter, 0 9 ; EPILOG: br i1 %lcmp.mod, label %for.body.epil.preheader, label %for.end.loopexit 10 ; EPILOG: for.body.epil: 11 12 ; PROLOG: %xtraiter = and i32 %n 13 ; PROLOG: %lcmp.mod = icmp ne i32 %xtraiter, 0 14 ; PROLOG: br i1 %lcmp.mod, label %for.body.prol.preheader, label %for.body.prol.loopexit 15 ; PROLOG: for.body.prol: 16 ; PROLOG: for.body: 17 18 define i32 @test(i32* nocapture %a, i32 %n) nounwind uwtable readonly { 19 entry: 20 %cmp1 = icmp eq i32 %n, 0 21 br i1 %cmp1, label %for.end, label %for.body 22 23 for.body: ; preds = %for.body, %entry 24 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ] 25 %sum.02 = phi i32 [ %add, %for.body ], [ 0, %entry ] 26 %arrayidx = getelementptr inbounds i32, i32* %a, i64 %indvars.iv 27 %0 = load i32, i32* %arrayidx, align 4 28 %add = add nsw i32 %0, %sum.02 29 %indvars.iv.next = add i64 %indvars.iv, 1 30 %lftr.wideiv = trunc i64 %indvars.iv.next to i32 31 %exitcond = icmp eq i32 %lftr.wideiv, %n 32 br i1 %exitcond, label %for.end, label %for.body 33 34 for.end: ; preds = %for.body, %entry 35 %sum.0.lcssa = phi i32 [ 0, %entry ], [ %add, %for.body ] 36 ret i32 %sum.0.lcssa 37 } 38 39 40
