Page MenuHomePhabricator

asbirlea (Alina Sbirlea)
Google

Projects

User does not belong to any projects.

User Details

User Since
Feb 10 2016, 9:51 AM (241 w, 6 d)

Recent Activity

Fri, Sep 25

asbirlea added a comment to D87869: [NewPM] Set -enable-npm-optnone to true by default.

Add to the documentation on NPM passes.
I meant to add about isRequired, but with mentioning the behavior for optnone functions. I didn't mean to document what optnone means.

Fri, Sep 25, 12:18 PM · Restricted Project
asbirlea added a comment to D87869: [NewPM] Set -enable-npm-optnone to true by default.

Could you add a few lines about this in the LLVM documentation?
We've had all the discussion in patches, so it may get unnoticed if it's not documented.

Fri, Sep 25, 12:07 PM · Restricted Project
asbirlea accepted D87957: [LoopReroll][NewPM] Port -loop-reroll to NPM.
Fri, Sep 25, 12:05 PM · Restricted Project

Wed, Sep 23

asbirlea accepted D88128: [NewPM][MSSA] Fix failures under NPM due to -enable-mssa-loop-dependency.
Wed, Sep 23, 1:09 PM · Restricted Project
asbirlea updated subscribers of D87625: [PruneEH][NewPM] Port prune-eh to NPM.

I think for prune-eh, the NPM inliner was intended to address the same cleanup, but I don't have all the context.
@chandlerc for more details.

Wed, Sep 23, 12:18 PM · Restricted Project
asbirlea added inline comments to D87957: [LoopReroll][NewPM] Port -loop-reroll to NPM.
Wed, Sep 23, 12:17 PM · Restricted Project

Tue, Sep 22

asbirlea added a comment to D87817: [LICM][Coroutine] Don't sink stores to coroutine suspend point..

Can you please extend the testing to cover at least some of these cases: multiple predecesssors, multiple exit blocks, multiple predecessors with a coroutine switch?

Tue, Sep 22, 12:46 PM · Restricted Project

Mon, Sep 21

asbirlea accepted D87857: [NewPM] Pin -lazy-branch-prob and -lazy-block-freq tests to legacy PM.

Thanks for all these!

Mon, Sep 21, 5:35 PM · Restricted Project
asbirlea accepted D87741: [Delinearization][NewPM] Port delinearization to NPM.
Mon, Sep 21, 5:33 PM · Restricted Project
asbirlea accepted D88058: [UnifyLoopExits] Pin tests with -unify-loop-exits to legacy PM.
Mon, Sep 21, 5:04 PM · Restricted Project
asbirlea added a comment to D87734: [UnifyLoopExits][NewPM] Port unify-loop-exits to NPM.

I don't think this should be ported to the NPM.

Mon, Sep 21, 2:50 PM · Restricted Project
asbirlea accepted D87932: [NewPM] Support -disable-simplify-libcall/-disable-builtin in NPM opt.
Mon, Sep 21, 2:45 PM · Restricted Project
asbirlea accepted D87945: [NewPM] Pin tests with -debug-pass to legacy PM.

For the NPM, there's -debug-pass-manager, so this would give you the corresponding coverage (with a different CHECK prefix). Some tests already have this.
Optional to add coverage for phi-values-usage.ll and alias-analysis-uses.ll, the others are already covered or are checking for Changed.

Mon, Sep 21, 12:49 PM · Restricted Project
asbirlea accepted D88014: [LoopUnswitch] Trivial simplification: remove trivial dead condition after unswitch.

I think in general the pass is not trying to do DCE too, but here it looks cheap to clean the dead code left by the conditional.
Please have MSSA updated though.

Mon, Sep 21, 12:37 PM · Restricted Project
asbirlea accepted D87798: [NewPM][CGSCC] Handle newly added functions in updateCGAndAnalysisManagerForPass.
Mon, Sep 21, 12:18 PM · Restricted Project
asbirlea added inline comments to D87957: [LoopReroll][NewPM] Port -loop-reroll to NPM.
Mon, Sep 21, 12:13 PM · Restricted Project
asbirlea accepted D87933: [DIE] Remove DeadInstEliminationPass.

This looks like a good cleanup. It's possible for DIE to be useful as a "light-weight" DCE, but that could also be achieved by adding a flag to DCE.
Unless there are folks that come forward that they need this for other reasons, I think it's good to clean the codebase.

Mon, Sep 21, 11:44 AM · Restricted Project

Fri, Sep 18

asbirlea added a comment to D87798: [NewPM][CGSCC] Handle newly added functions in updateCGAndAnalysisManagerForPass.

The general comment (I mentioned this offline too), is that this is much closer to the direction I'd see the changes going. By this, I mean, the APIs removed here seemed unnecessary, and inlining their functionality at the callsite made more sense.
This seems the right approach for coroutines and outlining. I'm not clear for inlining at this point, Im working to understand the flow better there.

Fri, Sep 18, 3:13 PM · Restricted Project
asbirlea accepted D87922: [test][InstrProf] Fix always_inline.ll under NPM.
Fri, Sep 18, 1:45 PM · Restricted Project

Thu, Sep 17

asbirlea accepted D87778: [MemorySSA] Be more conservative when traversing MemoryPhis..
Thu, Sep 17, 2:07 PM · Restricted Project
asbirlea added inline comments to D87778: [MemorySSA] Be more conservative when traversing MemoryPhis..
Thu, Sep 17, 12:30 PM · Restricted Project

Wed, Sep 16

asbirlea committed rG344a3d0bc0fb: [MemorySSA] Rename uses in blocks with Phis. (authored by asbirlea).
[MemorySSA] Rename uses in blocks with Phis.
Wed, Sep 16, 5:27 PM
asbirlea closed D87661: [MemorySSA] Rename uses in blocks with Phis..
Wed, Sep 16, 5:27 PM · Restricted Project
asbirlea added a comment to D87661: [MemorySSA] Rename uses in blocks with Phis..

Incorrect previous approach. The renaming needs to start at the block with the new def, but it needs to rename all blocks in the IDF, not just those with phis added.

Wed, Sep 16, 5:23 PM · Restricted Project
asbirlea retitled D87661: [MemorySSA] Rename uses in blocks with Phis. from [MemorySSA] Rename uses starting at defining access to [MemorySSA] Rename uses in blocks with Phis..
Wed, Sep 16, 5:22 PM · Restricted Project
asbirlea updated the diff for D87661: [MemorySSA] Rename uses in blocks with Phis..

Update to different approach.

Wed, Sep 16, 5:21 PM · Restricted Project
asbirlea added a comment to D87778: [MemorySSA] Be more conservative when traversing MemoryPhis..

I'm in favor of being conservative here and checking for only constant indexes on the geps.

Wed, Sep 16, 2:37 PM · Restricted Project
asbirlea updated the summary of D87661: [MemorySSA] Rename uses in blocks with Phis..
Wed, Sep 16, 12:13 AM · Restricted Project
asbirlea updated the diff for D87661: [MemorySSA] Rename uses in blocks with Phis..

Updated the printing in separate revision. Rebase.

Wed, Sep 16, 12:13 AM · Restricted Project
asbirlea accepted D87743: [NPM] Translate alias analysis into require<> as well.
Wed, Sep 16, 12:07 AM · Restricted Project

Tue, Sep 15

asbirlea committed rGd3d76039002c: [MemorySSA] Report unoptimized as None, not MayAlias. (authored by asbirlea).
[MemorySSA] Report unoptimized as None, not MayAlias.
Tue, Sep 15, 11:59 PM
asbirlea added a comment to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

I checked in a fix in https://reviews.llvm.org/rGfc8200633122, but I haven't yet verified it addresses all the failures reported.

Tue, Sep 15, 11:37 PM · Restricted Project, Restricted Project
asbirlea committed rGfc8200633122: [MemorySSA] Set MustDominate to true for PhiTranslation. (authored by asbirlea).
[MemorySSA] Set MustDominate to true for PhiTranslation.
Tue, Sep 15, 11:31 PM
asbirlea added a comment to D87623: [CGSCC][NewPM] Fix adding mutually recursive new functions.

We have quite a few "outlining" passes but most are module passes I suppose.
Assuming we want to work towards CGSCC passes, and maybe, maybe parallelize them at some point, we should allow proper updates of the CG.

I am reasonably certain we will soonish look at integrating functionality that will subsume hot-cold splitting into the Attributor. At that point we will do real "outlining". Other use cases are certainly not unreasonable. What I try to say is that CoroSplit is probably not the only user of outlining.

Tue, Sep 15, 7:17 PM · Restricted Project
asbirlea committed rG3b3ca5c989f9: Fix test after D86156. (authored by asbirlea).
Fix test after D86156.
Tue, Sep 15, 7:14 PM
asbirlea added a comment to D87623: [CGSCC][NewPM] Fix adding mutually recursive new functions.

I'd like to revisit this and open a discussion.

Sure. I assume we don't need to revert the patch but we can start a discussion with no pressure, right?

Yes, exactly. The status before this patch was broken anyway, no sense reverting. I'm looking to understand whether there are ways to improve the infrastructure/APIs.

The createNode is only called in addNewFunctionIntoSCC and addNewFunctionIntoRefSCC, where the former is only called in unit tests and the latter is only used by co-routines. I'm trying to understand what the expected changes in the CallGraph are for some of these methods, if this is specific to coroutines or there are other usecases and if we can get better testing.
Is registerOutlinedFunction used out-of tree?

Except the Attributor and coroutines there is no user of the CallGraphUpdater (I know of).
The Attributor does not yet use this function but it is not unreasonable to expect it will be used.
There is a good chance the "shallow wrapper" functionality should actually do so, though that feature is on its own not the most useful anyway.

@jdoerfert, @modocache: Could you help me with some context please? (I'm happy to use llvm-dev@ for the discussion too.)

I'm fine with LLVM-dev, TBH I need to talk about updating the LazyCallGraph soon anyway. I think the current &updateCGAndAnalysisManagerForCGSCCPass in conjunction with the ::run method don't allow you to delete more than one function from the SCC at a time, which is bad. I might simply use it wrong but whatever it is, I'm all for a discussion on how we expose new PM CGSCC updates :) [= how we properly replace/implement something like the CallGraphUpdater]

Tue, Sep 15, 7:02 PM · Restricted Project
asbirlea added a comment to D87623: [CGSCC][NewPM] Fix adding mutually recursive new functions.

I'd like to revisit this and open a discussion. The createNode is only called in addNewFunctionIntoSCC and addNewFunctionIntoRefSCC, where the former is only called in unit tests and the latter is only used by co-routines. I'm trying to understand what the expected changes in the CallGraph are for some of these methods, if this is specific to coroutines or there are other usecases and if we can get better testing.
Is registerOutlinedFunction used out-of tree?

Tue, Sep 15, 5:31 PM · Restricted Project
asbirlea added a comment to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

Right!
I believe the solution is to set the location size to unknown after a phi translation with the same ptr, but I need to properly verify that.

// include/llvm/Analysis/MemorySSA.h:1233
CurrentPair.second = Location.getWithNewSize(LocationSize::unknown());
return
Tue, Sep 15, 1:32 PM · Restricted Project, Restricted Project
asbirlea added a comment to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

Yes, the load should have the Phi as the defining access. I'm still looking into where this information should come from, but it's not hitting the phi translation.
Thank you for the revert, I'll update with the fix once I have it available.

Tue, Sep 15, 12:44 PM · Restricted Project, Restricted Project

Mon, Sep 14

asbirlea accepted D87658: [RegionInfo][NewPM] Fix RegionInfo tests to work under NPM.
Mon, Sep 14, 7:25 PM · Restricted Project
asbirlea accepted D87664: [NewPM][opt] Fix -globals-aa not being recognized as alias analysis in NPM.
Mon, Sep 14, 7:22 PM · Restricted Project
asbirlea accepted D87660: [PostDominators][NewPM] Fix tests to work under NPM.
Mon, Sep 14, 6:20 PM · Restricted Project
asbirlea requested review of D87661: [MemorySSA] Rename uses in blocks with Phis..
Mon, Sep 14, 6:18 PM · Restricted Project
asbirlea added a comment to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

I'd suggest reverting if the failures are blocking, but we do need a reproducer so it can be recommitted after a fix is in place.

Mon, Sep 14, 5:41 PM · Restricted Project, Restricted Project
asbirlea accepted D87551: [LICM] Make Loop ICM profile aware again.

Thank you for the patch!

Mon, Sep 14, 2:24 PM · Restricted Project
asbirlea accepted D87632: [PruneEH][NFC] Use CallGraphUpdater in PruneEH.

lg

Mon, Sep 14, 2:14 PM · Restricted Project
asbirlea added a comment to D87625: [PruneEH][NewPM] Port prune-eh to NPM.

I think for clarify this could be split into a patch that updates the current pass to use the CallGraphUpdater, and a patch that only adds the pass for the new pass manager.

Mon, Sep 14, 1:30 PM · Restricted Project
asbirlea accepted D86979: [docs][NewPM] Add docs for writing NPM passes.

Thank you adding this!

Mon, Sep 14, 12:56 PM · Restricted Project
asbirlea added inline comments to D84905: [MemorySSA] Restrict optimizations after a PhiTranslation..
Mon, Sep 14, 11:51 AM · Restricted Project
asbirlea added inline comments to rGc4f1b3144184: [MemorySSA] Make sure PerformedPhiTrans is updated for each visited def..
Mon, Sep 14, 8:40 AM

Thu, Sep 10

asbirlea accepted D87163: [DSE] Switch to MemorySSA-backed DSE by default..

I'm running additional testing in the background. I don't have anything holding this back at this point.
Either check in and if something shows up, I'll let you know to resolve or revert, or wait another 1-2 days for the testing I'm running to complete.
I'm ok with either option, and may even be inclined for the former, because there may be additional parties coming forward if issues arise, and I'd like these to come up earlier rather than later.

Thu, Sep 10, 2:08 PM · Restricted Project, Restricted Project
asbirlea accepted D87475: Use pragmas to work around MSVC x86_32 debug miscompile bug.

Thank you so much Reid for taking the time to look into this, discovering the msvc issue and proposing this work-around!

Thu, Sep 10, 12:56 PM · Restricted Project

Wed, Sep 9

asbirlea added inline comments to D87163: [DSE] Switch to MemorySSA-backed DSE by default..
Wed, Sep 9, 8:24 PM · Restricted Project, Restricted Project
asbirlea requested changes to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

I'm running into a crash with this, could you please hold off until tomorrow? I'm working on getting a reproducer.

Wed, Sep 9, 7:02 PM · Restricted Project, Restricted Project
asbirlea added a comment to D87387: [DSE] Handle masked stores.

Could you please prioritize figuring out why the test fails with DSE+MSSA? It's important to have this clarified before the flag flip.

Wed, Sep 9, 12:18 PM · Restricted Project
asbirlea added a comment to D87322: [Passes] Move ADCE before DSE & LICM..

I forgot to add, if you could include the impact on the NPM for compile-time, it would be great.

Wed, Sep 9, 11:50 AM · Restricted Project
asbirlea accepted D87355: [LoopLoadElim] Filter away candidates that stop being AddRecs after loop versioning. PR47457.
Wed, Sep 9, 11:36 AM · Restricted Project
asbirlea added a comment to D87322: [Passes] Move ADCE before DSE & LICM..

It may be good to have this in before the DSE+MSSA flag flip to minimize the regression introduced by that patch (assuming there are no meaningful binary changes or run time regressions for the NPM either).

Wed, Sep 9, 11:29 AM · Restricted Project
asbirlea accepted D87163: [DSE] Switch to MemorySSA-backed DSE by default..

Green light from me as well. I think there's still room for improvement for the NPM, but this is the right step to move forward infrastructure-wise.

Wed, Sep 9, 11:23 AM · Restricted Project, Restricted Project
asbirlea accepted D87386: [DSE,MemorySSA] Handle atomic stores expliclty in isReadClobber..

lgtm

Wed, Sep 9, 11:19 AM · Restricted Project
asbirlea accepted D87387: [DSE] Handle masked stores.

lgtm

Wed, Sep 9, 11:16 AM · Restricted Project

Tue, Sep 8

asbirlea accepted D87137: [EarlyCSE] Explicitly require AAResultsWrapperPass..
Tue, Sep 8, 5:08 PM · Restricted Project
asbirlea accepted D87061: Handle masked loads and stores in MemoryLocation/Dependence.

This looks good to me.

Tue, Sep 8, 5:05 PM · Restricted Project
asbirlea added inline comments to D87293: [GVN] Fix undef incoming value for phi node when new loop exit block created..
Tue, Sep 8, 3:59 PM · Restricted Project
asbirlea added a comment to D87322: [Passes] Move ADCE before DSE & LICM..

Did you investigate a similar change for lib/Passes/PassBuilder.cpp:734->720?

Tue, Sep 8, 3:51 PM · Restricted Project
asbirlea accepted D87319: [DomTree] Use SmallVector<DomTreeNodeBase *, 4> instead of std::vector..

LGTM.

Tue, Sep 8, 12:21 PM · Restricted Project

Thu, Sep 3

asbirlea accepted D86651: [MemCpyOpt] Preserve MemorySSA..

Diff looks good to me.
Same comment as before about testing with EXPENSIVE_CHECKS to catch potential missed updates.
Thank you for working on this!

Thu, Sep 3, 3:51 PM · Restricted Project

Wed, Sep 2

asbirlea accepted D86534: [GVN] Preserve MemorySSA if it is available..

Changes look good.
Please also test with EXPENSIVE_CHECKS. The additional MSSA verifications there can unearth missed changes.

Wed, Sep 2, 6:55 PM · Restricted Project
asbirlea accepted D86967: [PassManager] Move load/store motion pass after DSE in LTO pipeline..

Unless there are complaints from others, I think this is good.

Wed, Sep 2, 4:42 PM · Restricted Project
asbirlea added a comment to D86651: [MemCpyOpt] Preserve MemorySSA..

A small note to test with EXPENSIVE_CHECKS as well, as there's extra verification there.

Wed, Sep 2, 4:32 PM · Restricted Project
asbirlea accepted D87027: [MemCpyOptimizer] Change required analysis order for BasicAA/PhiValuesAnalysis.

I realized I accidentally reordered passes, but didn't realize the impact of not using phi values in basicaa.

Wed, Sep 2, 12:11 PM · Restricted Project

Tue, Sep 1

asbirlea added a comment to D74640: [LoopRotate] Add explicit flag to require MSSA..

Thank you for the results! Looking at libclamav_unsp.c, I'm seeing 3.96% in instructions for ThinLTO, which I'm guessing matches the results you see.
The fluctuations may make sense, since with the additional MSSA updates in LoopRotate it can do more or less work, depending on the updates. For instructions per cycle I'm seeing it neutral results (0.5%), and for wall time (average over 20 run), also neutral.

Tue, Sep 1, 9:06 PM · Restricted Project
asbirlea committed rGce66089ac6f0: Fix build-bots. (authored by asbirlea).
Fix build-bots.
Tue, Sep 1, 8:25 PM
asbirlea added a comment to D86651: [MemCpyOpt] Preserve MemorySSA..

I was working on having MemCpyOpt preserve and use MSSA, in order to replace the use of MemDepAnalysis entirely. But I'm happy with having MSSA preserved first :-).

Tue, Sep 1, 5:50 PM · Restricted Project
asbirlea committed rG1ccfb52a6174: [MemCpyOptimizer] Preserve analyses and replace use of lambdas to get them. (authored by asbirlea).
[MemCpyOptimizer] Preserve analyses and replace use of lambdas to get them.
Tue, Sep 1, 5:37 PM
asbirlea closed D74494: [MemCpyOptimizer] Preserve analyses and replace use of lambdas to get them..
Tue, Sep 1, 5:36 PM · Restricted Project
asbirlea updated the diff for D74494: [MemCpyOptimizer] Preserve analyses and replace use of lambdas to get them..

Rebase, update tests.

Tue, Sep 1, 5:35 PM · Restricted Project
asbirlea added a reviewer for D74494: [MemCpyOptimizer] Preserve analyses and replace use of lambdas to get them.: fhahn.
Tue, Sep 1, 3:37 PM · Restricted Project
asbirlea added a comment to D86967: [PassManager] Move load/store motion pass after DSE in LTO pipeline..

The idea to have all passes in sequence preserved and use MemorySSA is great. I'm wondering if changing the pass order affects run time (due to potential missed optimizations).
On the other hand, this only affects the LPM, so it may not be as critical due to the plan to switch to the NPM.

Tue, Sep 1, 3:02 PM · Restricted Project
asbirlea added a comment to D74640: [LoopRotate] Add explicit flag to require MSSA..

Rebased after the fix to remove DT recomputation in MSSA. I'm still seeing a small regression on one of the tests in the original PRs (0.6% in instructions on mogrify.i), but it's nowhere near the regression that motivated the pipeline split.
I'm not sure if it's worth pushing this forward if the only motivation is re-merging the pipeline in the LPM, unless it's entirely performance neutral.

Tue, Sep 1, 2:05 PM · Restricted Project
asbirlea updated the diff for D74640: [LoopRotate] Add explicit flag to require MSSA..

Rebase at ToT.

Tue, Sep 1, 1:58 PM · Restricted Project
asbirlea committed rGc292fba46fc2: [MemorySSA] Update phi map with replacement value. (authored by asbirlea).
[MemorySSA] Update phi map with replacement value.
Tue, Sep 1, 11:58 AM

Mon, Aug 31

asbirlea committed rG63844c116a0a: [MemorySSA] Clean up single value phis. (authored by asbirlea).
[MemorySSA] Clean up single value phis.
Mon, Aug 31, 7:27 PM

Aug 28 2020

asbirlea accepted D85676: [NewPM] collapsing nested pass mangers of the same type.

Seems like a good idea. I'm curious if you saw any impact from this change.

Aug 28 2020, 5:35 PM · Restricted Project, Restricted Project
asbirlea accepted D86526: [SimplifyCFG] Consider cost of combining predicates..

Please clang-format

Aug 28 2020, 5:32 PM · Restricted Project
asbirlea added inline comments to D86534: [GVN] Preserve MemorySSA if it is available..
Aug 28 2020, 5:12 PM · Restricted Project

Aug 27 2020

asbirlea committed rGd370836c20c5: [MemorySSA] Assert defining access is not a MemoryUse. (authored by asbirlea).
[MemorySSA] Assert defining access is not a MemoryUse.
Aug 27 2020, 6:22 PM
asbirlea accepted D85561: [BasicBlockUtils] Factor MergeBlockIntoPredecessor args into a struct, NFC.

lgtm with some nits.

Aug 27 2020, 5:32 PM · Restricted Project
asbirlea added a comment to D86534: [GVN] Preserve MemorySSA if it is available..

Not exactly the path I had in mind (working on newGVN was), but that's a longer avenue and I'm curious to see the compile-time impact of preserving MSSA :-)

Aug 27 2020, 4:37 PM · Restricted Project
asbirlea accepted D86156: [BFI] Make BFI information available through loop passes inside LoopStandardAnalysisResults.

Diff looks reasonable at this point. Thank you for the patch!
Please wait on @nikic for compile-time impact or additional feedback.

Aug 27 2020, 4:16 PM · Restricted Project, Restricted Project
asbirlea accepted D86487: [DSE,MemorySSA] Check if DomAccess is valid for elimination first..

A few nits, but lgtm.

Aug 27 2020, 3:58 PM · Restricted Project
asbirlea added inline comments to D86156: [BFI] Make BFI information available through loop passes inside LoopStandardAnalysisResults.
Aug 27 2020, 1:55 PM · Restricted Project, Restricted Project
asbirlea added a comment to D86156: [BFI] Make BFI information available through loop passes inside LoopStandardAnalysisResults.

Your understanding is correct. I only have one more comment on preserving passes in LICM too.

Aug 27 2020, 12:22 PM · Restricted Project, Restricted Project

Aug 26 2020

asbirlea accepted D86663: [test][Inliner] Make always-inline.ll work with NPM.
Aug 26 2020, 5:04 PM · Restricted Project
asbirlea added inline comments to D86156: [BFI] Make BFI information available through loop passes inside LoopStandardAnalysisResults.
Aug 26 2020, 4:54 PM · Restricted Project, Restricted Project
asbirlea accepted D86486: [DSE,MemorySSA] Limit walker steps..
Aug 26 2020, 3:52 PM · Restricted Project
asbirlea committed rG0b3422630434: Use properlyDominates in RDFLiveness when sorting on dominance. (authored by asbirlea).
Use properlyDominates in RDFLiveness when sorting on dominance.
Aug 26 2020, 3:17 PM
asbirlea closed D86661: Use properlyDominates in RDFLiveness when sorting on dominance..
Aug 26 2020, 3:17 PM · Restricted Project
asbirlea accepted D86661: Use properlyDominates in RDFLiveness when sorting on dominance..
Aug 26 2020, 2:16 PM · Restricted Project

Aug 25 2020

asbirlea added inline comments to D86486: [DSE,MemorySSA] Limit walker steps..
Aug 25 2020, 2:32 PM · Restricted Project