<div dir="ltr">Our Android and Windows bots are unhappy with these changes: <a href="http://lab.llvm.org:8011/builders/sanitizer-windows/builds/27045">http://lab.llvm.org:8011/builders/sanitizer-windows/builds/27045</a> and <a href="http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-android/builds/9491">http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-android/builds/9491</a><br></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Apr 9, 2018 at 4:37 PM, Michael Zolotukhin via llvm-commits <span dir="ltr"><<a href="mailto:llvm-commits@lists.llvm.org" target="_blank">llvm-commits@lists.llvm.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Author: mzolotukhin<br>
Date: Mon Apr 9 16:37:37 2018<br>
New Revision: 329644<br>
<br>
URL: <a href="http://llvm.org/viewvc/llvm-project?rev=329644&view=rev" rel="noreferrer" target="_blank">http://llvm.org/viewvc/llvm-<wbr>project?rev=329644&view=rev</a><br>
Log:<br>
[PR16756] Use SSAUpdaterBulk in JumpThreading.<br>
<br>
Summary:<br>
SSAUpdater is a bottleneck in JumpThreading, and this patch improves the<br>
situation by using SSAUpdaterBulk instead.<br>
<br>
Compile time impact: no noticable changes on CTMark, a big improvement<br>
on the test from PR16756.<br>
<br>
Reviewers: dberlin, davide, MatzeB<br>
<br>
Subscribers: llvm-commits, hiraditya<br>
<br>
Differential Revision: <a href="https://reviews.llvm.org/D44282" rel="noreferrer" target="_blank">https://reviews.llvm.org/<wbr>D44282</a><br>
<br>
Modified:<br>
llvm/trunk/lib/Transforms/<wbr>Scalar/JumpThreading.cpp<br>
<br>
Modified: llvm/trunk/lib/Transforms/<wbr>Scalar/JumpThreading.cpp<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/JumpThreading.cpp?rev=329644&r1=329643&r2=329644&view=diff" rel="noreferrer" target="_blank">http://llvm.org/viewvc/llvm-<wbr>project/llvm/trunk/lib/<wbr>Transforms/Scalar/<wbr>JumpThreading.cpp?rev=329644&<wbr>r1=329643&r2=329644&view=diff</a><br>
==============================<wbr>==============================<wbr>==================<br>
--- llvm/trunk/lib/Transforms/<wbr>Scalar/JumpThreading.cpp (original)<br>
+++ llvm/trunk/lib/Transforms/<wbr>Scalar/JumpThreading.cpp Mon Apr 9 16:37:37 2018<br>
@@ -66,6 +66,7 @@<br>
#include "llvm/Transforms/Utils/<wbr>BasicBlockUtils.h"<br>
#include "llvm/Transforms/Utils/<wbr>Cloning.h"<br>
#include "llvm/Transforms/Utils/<wbr>SSAUpdater.h"<br>
+#include "llvm/Transforms/Utils/<wbr>SSAUpdaterBulk.h"<br>
#include "llvm/Transforms/Utils/<wbr>ValueMapper.h"<br>
#include <algorithm><br>
#include <cassert><br>
@@ -1989,9 +1990,13 @@ bool JumpThreadingPass::ThreadEdge(<wbr>Basic<br>
// now have to update all uses of the value to use either the original value,<br>
// the cloned value, or some PHI derived value. This can require arbitrary<br>
// PHI insertion, of which we are prepared to do, clean these up now.<br>
- SSAUpdater SSAUpdate;<br>
+ SSAUpdaterBulk SSAUpdate;<br>
SmallVector<Use*, 16> UsesToRename;<br>
+<br>
+ unsigned VarNum = 0;<br>
for (Instruction &I : *BB) {<br>
+ UsesToRename.clear();<br>
+<br>
// Scan all uses of this instruction to see if it is used outside of its<br>
// block, and if so, record them in UsesToRename.<br>
for (Use &U : I.uses()) {<br>
@@ -2008,19 +2013,15 @@ bool JumpThreadingPass::ThreadEdge(<wbr>Basic<br>
// If there are no uses outside the block, we're done with this instruction.<br>
if (UsesToRename.empty())<br>
continue;<br>
+ SSAUpdate.AddVariable(VarNum, I.getName(), I.getType());<br>
<br>
- DEBUG(dbgs() << "JT: Renaming non-local uses of: " << I << "\n");<br>
-<br>
- // We found a use of I outside of BB. Rename all uses of I that are outside<br>
- // its block to be uses of the appropriate PHI node etc. See ValuesInBlocks<br>
- // with the two values we know.<br>
- SSAUpdate.Initialize(I.<wbr>getType(), I.getName());<br>
- SSAUpdate.AddAvailableValue(<wbr>BB, &I);<br>
- SSAUpdate.AddAvailableValue(<wbr>NewBB, ValueMapping[&I]);<br>
-<br>
- while (!UsesToRename.empty())<br>
- SSAUpdate.RewriteUse(*<wbr>UsesToRename.pop_back_val());<br>
- DEBUG(dbgs() << "\n");<br>
+ // We found a use of I outside of BB - we need to rename all uses of I that<br>
+ // are outside its block to be uses of the appropriate PHI node etc.<br>
+ SSAUpdate.AddAvailableValue(<wbr>VarNum, BB, &I);<br>
+ SSAUpdate.AddAvailableValue(<wbr>VarNum, NewBB, ValueMapping[&I]);<br>
+ for (auto U : UsesToRename)<br>
+ SSAUpdate.AddUse(VarNum, U);<br>
+ VarNum++;<br>
}<br>
<br>
// Ok, NewBB is good to go. Update the terminator of PredBB to jump to<br>
@@ -2037,6 +2038,10 @@ bool JumpThreadingPass::ThreadEdge(<wbr>Basic<br>
{DominatorTree::Insert, PredBB, NewBB},<br>
{DominatorTree::Delete, PredBB, BB}});<br>
<br>
+ // Apply all updates we queued with DDT and get the updated Dominator Tree.<br>
+ DominatorTree *DT = &DDT->flush();<br>
+ SSAUpdate.RewriteAllUses(DT);<br>
+<br>
// At this point, the IR is fully up to date and consistent. Do a quick scan<br>
// over the new instructions and zap any that are constants or dead. This<br>
// frequently happens because of phi translation.<br>
<br>
<br>
______________________________<wbr>_________________<br>
llvm-commits mailing list<br>
<a href="mailto:llvm-commits@lists.llvm.org">llvm-commits@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits" rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/llvm-commits</a><br>
</blockquote></div><br></div>