diff options
| author | Michael Zolotukhin <mzolotukhin@apple.com> | 2014-11-20 20:19:55 +0000 |
|---|---|---|
| committer | Michael Zolotukhin <mzolotukhin@apple.com> | 2014-11-20 20:19:55 +0000 |
| commit | 4e7b10b07f0ca2c8185e96bdc58ae4d9d540b3a4 (patch) | |
| tree | daf02aed512ca1f354e9af1671ffb7545e99129f /test/Transforms | |
| parent | d23f04a165d7682fc09f5944cac66a7895e48dfe (diff) | |
Fix a trip-count overflow issue in LoopUnroll.
Currently LoopUnroll generates a prologue loop before the main loop
body to execute first N%UnrollFactor iterations. Also, this loop is
used if trip-count can overflow - it's determined by a runtime check.
However, we've been mistakenly optimizing this loop to a linear code for
UnrollFactor = 2, not taking into account that it also serves as a safe
version of the loop if its trip-count overflows.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@222451 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/Transforms')
| -rw-r--r-- | test/Transforms/LoopUnroll/runtime-loop1.ll | 2 | ||||
| -rw-r--r-- | test/Transforms/LoopUnroll/tripcount-overflow.ll | 30 |
2 files changed, 31 insertions, 1 deletions
diff --git a/test/Transforms/LoopUnroll/runtime-loop1.ll b/test/Transforms/LoopUnroll/runtime-loop1.ll index 5ff75e33f7f..38b4f32354a 100644 --- a/test/Transforms/LoopUnroll/runtime-loop1.ll +++ b/test/Transforms/LoopUnroll/runtime-loop1.ll @@ -3,7 +3,7 @@ ; This tests that setting the unroll count works ; CHECK: for.body.prol: -; CHECK: br label %for.body.preheader.split +; CHECK: br i1 %prol.iter.cmp, label %for.body.prol, label %for.body.preheader.split ; CHECK: for.body: ; CHECK: br i1 %exitcond.1, label %for.end.loopexit.unr-lcssa, label %for.body ; CHECK-NOT: br i1 %exitcond.4, label %for.end.loopexit{{.*}}, label %for.body diff --git a/test/Transforms/LoopUnroll/tripcount-overflow.ll b/test/Transforms/LoopUnroll/tripcount-overflow.ll new file mode 100644 index 00000000000..d59368578ec --- /dev/null +++ b/test/Transforms/LoopUnroll/tripcount-overflow.ll @@ -0,0 +1,30 @@ +; RUN: opt < %s -S -unroll-runtime -unroll-count=2 -loop-unroll | FileCheck %s +target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128" + +; When prologue is fully unrolled, the branch on its end is unconditional. +; Unrolling it is illegal if we can't prove that trip-count+1 doesn't overflow, +; like in this example, where it comes from an argument. +; +; This test is based on an example from here: +; http://stackoverflow.com/questions/23838661/why-is-clang-optimizing-this-code-out +; +; CHECK: while.body.prol: +; CHECK: br i1 +; CHECK: entry.split: + +; Function Attrs: nounwind readnone ssp uwtable +define i32 @foo(i32 %N) #0 { +entry: + br label %while.body + +while.body: ; preds = %while.body, %entry + %i = phi i32 [ 0, %entry ], [ %inc, %while.body ] + %cmp = icmp eq i32 %i, %N + %inc = add i32 %i, 1 + br i1 %cmp, label %while.end, label %while.body + +while.end: ; preds = %while.body + ret i32 %i +} + +attributes #0 = { nounwind readnone ssp uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" } |