Page MenuHomePhabricator

markus (Markus Lavin)
User

Projects

User does not belong to any projects.

User Details

User Since
Nov 19 2018, 1:12 AM (189 w, 2 d)

Recent Activity

Mon, Jun 20

markus committed rG3815ae29b5cb: [machinesink] fix debug invariance issue (authored by markus).
[machinesink] fix debug invariance issue
Mon, Jun 20, 11:13 PM · Restricted Project, Restricted Project
markus closed D127208: MachineSink debug invariance.
Mon, Jun 20, 11:13 PM · Restricted Project, Restricted Project

Mon, Jun 13

markus added a comment to D127208: MachineSink debug invariance.

ping

Mon, Jun 13, 9:37 PM · Restricted Project, Restricted Project

Fri, Jun 10

markus updated the diff for D127208: MachineSink debug invariance.

MachineBasicBlock::sizeWithoutDebugLargerThan(unsigned Limit)

Fri, Jun 10, 1:57 AM · Restricted Project, Restricted Project

Thu, Jun 9

markus updated the diff for D127208: MachineSink debug invariance.

Fixed typo in lit-test.

Thu, Jun 9, 6:34 AM · Restricted Project, Restricted Project
markus added inline comments to D127208: MachineSink debug invariance.
Thu, Jun 9, 2:43 AM · Restricted Project, Restricted Project
markus updated the diff for D127208: MachineSink debug invariance.

Added lit-test.

Thu, Jun 9, 2:39 AM · Restricted Project, Restricted Project

Tue, Jun 7

markus added inline comments to D127208: MachineSink debug invariance.
Tue, Jun 7, 7:03 AM · Restricted Project, Restricted Project
markus updated the diff for D127208: MachineSink debug invariance.

Add MachineBasicBlock::sizeWithoutDebug() to make more similar to IR.

Tue, Jun 7, 6:59 AM · Restricted Project, Restricted Project
markus added a comment to D127208: MachineSink debug invariance.

Should add a .mir test

Tue, Jun 7, 5:29 AM · Restricted Project, Restricted Project
markus requested review of D127208: MachineSink debug invariance.
Tue, Jun 7, 5:27 AM · Restricted Project, Restricted Project

May 24 2022

markus committed rGbb8e02325ffc: llvm-reduce: improve basic-blocks removal pass (authored by markus).
llvm-reduce: improve basic-blocks removal pass
May 24 2022, 12:54 AM · Restricted Project, Restricted Project
markus closed D125766: [llvm-reduce] improve bb removal.
May 24 2022, 12:53 AM · Restricted Project, Restricted Project

May 23 2022

markus added a comment to D125766: [llvm-reduce] improve bb removal.

I’m working on the machine equivalent of this reduction now and may make additional changes here

May 23 2022, 10:59 PM · Restricted Project, Restricted Project
markus added a comment to D125766: [llvm-reduce] improve bb removal.

@arsenm, is this what you had in mind wrt phi-nodes or were you requesting some more elaborate change?

May 23 2022, 6:01 AM · Restricted Project, Restricted Project

May 19 2022

markus updated the diff for D125766: [llvm-reduce] improve bb removal.

Do not branch to blocks with existing phi-node.

May 19 2022, 2:25 AM · Restricted Project, Restricted Project

May 18 2022

markus added a comment to D125766: [llvm-reduce] improve bb removal.

If you blindly scan forward, you're going to run into blocks that have phis that are now missing an input for the new predecessor

May 18 2022, 11:16 PM · Restricted Project, Restricted Project

May 17 2022

markus requested review of D125766: [llvm-reduce] improve bb removal.
May 17 2022, 5:13 AM · Restricted Project, Restricted Project

May 3 2022

markus committed rGdd8cf372c54f: [NFC] Minimal refactor of TTI to avoid clangsa complaint (authored by markus).
[NFC] Minimal refactor of TTI to avoid clangsa complaint
May 3 2022, 1:48 AM · Restricted Project, Restricted Project
markus closed D124754: Minimal refactor of TargetTransformInfoImpl.h to avoid Clang Static Analyser's core.CallAndMessage (NFC).
May 3 2022, 1:47 AM · Restricted Project, Restricted Project

May 2 2022

markus requested review of D124754: Minimal refactor of TargetTransformInfoImpl.h to avoid Clang Static Analyser's core.CallAndMessage (NFC).
May 2 2022, 2:21 AM · Restricted Project, Restricted Project

Apr 27 2022

markus accepted D124049: llvm-reduce: Introduce new scoring mechanism for MIR reductions.
Apr 27 2022, 11:57 PM · Restricted Project, Restricted Project

Apr 25 2022

markus accepted D124383: llvm-reduce: Do not try to delete frame instructions.
Apr 25 2022, 10:50 PM · Restricted Project, Restricted Project
markus added inline comments to D124383: llvm-reduce: Do not try to delete frame instructions.
Apr 25 2022, 7:37 AM · Restricted Project, Restricted Project
markus added a comment to D124383: llvm-reduce: Do not try to delete frame instructions.

I recall running into this as well. So this patch will make sure that they are never deleted thus improving reduction speed. But wasn't there a chance that the oracle would indicated that both instructions of the pair were to be deleted which could result in a valid reduction? I guess this is fine though and a better solution would be to deal with the frame pairs in a separate reduction pass anyway.

Apr 25 2022, 7:00 AM · Restricted Project, Restricted Project

Apr 24 2022

markus added a comment to D124008: llvm-reduce: Support multiple MachineFunctions.

Looks fine to me though someone with better insight into MMI could probably provide better feedback.

Apr 24 2022, 11:51 PM · Restricted Project, Restricted Project
markus added a comment to D124049: llvm-reduce: Introduce new scoring mechanism for MIR reductions.

Right, so with MIR reductions we totally forgot about this an relied on the IR serialization size which is obviously wrong. This looks like an improvement to me. Of course the details of the metric can always be discussed but this should be better than nothing. Especially if we start inserting IMPLICIT_DEF instructions during reduction.

Apr 24 2022, 11:13 PM · Restricted Project, Restricted Project
markus added a comment to D123953: llvm-reduce: Add cloning of target MachineFunctionInfo.

Having to touch all these target specific files seems a bit unfortunate but maybe that is the way it has to be.

Apr 24 2022, 11:01 PM · Restricted Project, Restricted Project
markus accepted D124050: llvm-reduce: Fix not removing first instruction in MachineBasicBlock.
Apr 24 2022, 10:56 PM · Restricted Project, Restricted Project
markus accepted D124034: llvm-reduce: Fix mangling types of generic registers.
Apr 24 2022, 10:51 PM · Restricted Project, Restricted Project
markus accepted D123935: llvm-reduce: Preserve frame index values when cloning function.

In general I am not too familiar with the frame index stuff so I cannot really comment much on that besides that it looks reasonable. From the point of view of the guy who did the initial MIR support work for llvm-reduce it looks like a good improvement and a step in the right direction.

Apr 24 2022, 10:41 PM · Restricted Project, Restricted Project

Apr 22 2022

markus added a comment to D123300: [Clang] Enable opaque pointers by default.

Sounds reasonable in general -- though isn't this a pre-existing problem you'd see if you simply had multiple loads from the same global (without any GEP)? Looking at the current ConstantHoist code, it doesn't seem to really consider the case where a global symbol address can be expensive, it only handles hoisting of large integer values.

Apr 22 2022, 6:51 AM · Restricted Project, Restricted Project, Restricted Project
markus added a comment to D123300: [Clang] Enable opaque pointers by default.

We have run into a slight performance degrading issue with our downstream target. The situation is that we relay on the "consthoist" pass with option "-consthoist-gep=1" to factor out the common parts of GEP expresseions to reduce the number of symbol references.
For example with the old typed pointers we hade just before that pass

store i16 100, i16* getelementptr inbounds ([8 x i16], [8 x i16]* @extUeValidatedTps, i16 0, i16 0), align 1
store i16 101, i16* getelementptr inbounds ([8 x i16], [8 x i16]* @extUeValidatedTps, i16 0, i16 1), align 1
store i16 102, i16* getelementptr inbounds ([8 x i16], [8 x i16]* @extUeValidatedTps, i16 0, i16 2), align 1

resulting in that the first GEP was factored out (with the other two based on it) so in total we had one symbol reference.

Apr 22 2022, 5:05 AM · Restricted Project, Restricted Project, Restricted Project

Apr 20 2022

markus added a comment to D124097: llvm-reduce: Add reduction pass to remove regalloc hints.

As a general thought, I wonder if it would be wise to keep the MIR specific reducer passes separate from the IR ones. Either by directory or file naming conventions.

Apr 20 2022, 11:09 PM · Restricted Project, Restricted Project
markus added inline comments to D124050: llvm-reduce: Fix not removing first instruction in MachineBasicBlock.
Apr 20 2022, 5:33 AM · Restricted Project, Restricted Project
markus accepted D123998: llvm-reduce: Try to parse triple/datalayout from module.

LGTM

Apr 20 2022, 5:29 AM · Restricted Project, Restricted Project
markus accepted D124005: llvm-reduce: Preserve subregisters and other fields for top block def.

Have not looked very closely at the tests but the code changes LGTM.

Apr 20 2022, 5:22 AM · Restricted Project, Restricted Project
markus added a comment to D123300: [Clang] Enable opaque pointers by default.

@markus Without tracing through it in detail, I'd guess that without opaque pointers this creates two getelementptr constant expressions that get folded together. With opaque pointers, the first one (which would be a zero-index GEP) is omitted, and only the second one is left. ConstantFolding will later canonicalize the GEP source type, but this only happens when InstCombine runs, while you're looking at unoptimized IR.

Apr 20 2022, 1:31 AM · Restricted Project, Restricted Project, Restricted Project

Apr 19 2022

markus added a comment to D123300: [Clang] Enable opaque pointers by default.

With

$ ./bin/clang -cc1 -emit-llvm BBI-68673.c

the IR contains the following

@a2_ldm_i64 = global [4 x [18 x { i64, i64 }]] zeroinitializer, align 16
%arrayidx = getelementptr inbounds [4 x [18 x { i64, i64 }]], ptr @a2_ldm_i64, i64 0, i64 %idxprom
%.real = load volatile i64, ptr getelementptr inbounds ([18 x { i64, i64 }], ptr @a2_ldm_i64, i64 0, i64 1), align 16

but I would expect both GEPs to have the same first argument type (as the global symbol)

Apr 19 2022, 7:25 AM · Restricted Project, Restricted Project, Restricted Project
markus added a reviewer for D123998: llvm-reduce: Try to parse triple/datalayout from module: reames.

I think @reames requested this back in D110527

Apr 19 2022, 6:25 AM · Restricted Project, Restricted Project

Apr 14 2022

markus accepted D123695: llvm-reduce: Fix handling of generic virtual registers.
Apr 14 2022, 12:32 AM · Restricted Project, Restricted Project
markus added a comment to D123695: llvm-reduce: Fix handling of generic virtual registers.

LGTM and happy to see that there is interest in llvm-reduce MIR reduction.

Apr 14 2022, 12:32 AM · Restricted Project, Restricted Project

Feb 2 2022

markus committed rG0d36d84de5f8: [llvm-reduce] Display all relevant options in -help (authored by markus).
[llvm-reduce] Display all relevant options in -help
Feb 2 2022, 12:46 AM
markus closed D118682: [llvm-reduce] Make all llvm-reduce specifict options use the same options category.
Feb 2 2022, 12:46 AM · Restricted Project
markus abandoned D104628: Fix a not debug invariant issue in GlobalOpt/GlobalStatus.

This fix and underlying issue has been superseded and resolved by of https://github.com/llvm/llvm-project/commit/236fbf571dc6cebcb81ac5187a170c8de6d5bc0e

Feb 2 2022, 12:27 AM · debug-info, Restricted Project

Feb 1 2022

markus requested review of D118682: [llvm-reduce] Make all llvm-reduce specifict options use the same options category.
Feb 1 2022, 12:47 AM · Restricted Project
markus committed rG8a8af120283d: [llvm-reduce] Set ShouldPreserveUseListOrder=true (authored by markus).
[llvm-reduce] Set ShouldPreserveUseListOrder=true
Feb 1 2022, 12:25 AM
markus closed D118585: [llvm-reduce] Specify ShouldPreserveUseListOrder for internal .ll dumps.
Feb 1 2022, 12:25 AM · Restricted Project

Jan 31 2022

markus added a comment to D118585: [llvm-reduce] Specify ShouldPreserveUseListOrder for internal .ll dumps.

This makes sense to me, but I think this is another motivation to use bitcode as default for temporary files (flip the default of TmpFilesAsBitcode to true)

Jan 31 2022, 2:45 AM · Restricted Project
markus requested review of D118585: [llvm-reduce] Specify ShouldPreserveUseListOrder for internal .ll dumps.
Jan 31 2022, 2:03 AM · Restricted Project

Jan 17 2022

markus added a comment to D117184: Fix failing assertion in SimplifyCFG.cpp as icmp gep fold to constant expression.

For some reason, the pipeline does not get reduced further (cc @markus) even though we can run the first part (the cgscc pipeline) and then feed it into the second part (the function pipeline).

Just looked into this a bit. Apparently the pipeline reduction looks into nested pass managers and will split up the cgscc pass manager, doing something like cgscc(inline) and cgscc(the-rest),function(simplifycfg). This no longer reproduces the issue, and the reduction is aborted early.

I think that either the early abort should be dropped so we later still try the cgscc(inline,the-rest) and function(simplifycfg) split, or pass manager splitting should be an entirely separate reduction step.

Jan 17 2022, 3:25 AM · Restricted Project
markus added a comment to D117184: Fix failing assertion in SimplifyCFG.cpp as icmp gep fold to constant expression.

For some reason, the pipeline does not get reduced further (cc @markus) even though we can run the first part (the cgscc pipeline) and then feed it into the second part (the function pipeline).

Thanks for letting me know @nikic. I will have a look to see what could be the problem and provide a fix if possible.

Jan 17 2022, 2:28 AM · Restricted Project

Nov 30 2021

markus committed rGce22b7f17b6a: [NPM] Fix LoopNestPasses in -print-pipeline-passes (authored by markus).
[NPM] Fix LoopNestPasses in -print-pipeline-passes
Nov 30 2021, 10:59 PM
markus closed D114771: [NPM] -print-pipeline-passes fix printing of LoopNestPasses.
Nov 30 2021, 10:58 PM · Restricted Project

Nov 29 2021

markus requested review of D114771: [NPM] -print-pipeline-passes fix printing of LoopNestPasses.
Nov 29 2021, 11:19 PM · Restricted Project

Nov 11 2021

markus committed rG4e94e25c9024: Fix minor deficiency in machine-sink. (authored by markus).
Fix minor deficiency in machine-sink.
Nov 11 2021, 11:16 PM
markus closed D111531: Fix minor deficiency in MachineSink..
Nov 11 2021, 11:16 PM · Restricted Project
markus added inline comments to D111531: Fix minor deficiency in MachineSink..
Nov 11 2021, 5:55 AM · Restricted Project
markus added inline comments to D111531: Fix minor deficiency in MachineSink..
Nov 11 2021, 5:03 AM · Restricted Project
markus edited reviewers for D111531: Fix minor deficiency in MachineSink., added: SjoerdMeijer, kazu, StephenTozer; removed: ddunbar, dexonsmith, dsanders.
Nov 11 2021, 2:06 AM · Restricted Project
markus added a comment to D110527: [llvm-reduce] Add MIR support..

The IR part can often be eliminated from a MIR based test case. But sure, one could argue that when running llvm-reduce it could be mandatory to add an IR section in the .mir file first (but then I think one needs to also add the function prototype).

I can maybe see this as an argument for allowing override of the triple, but not as an argument for not defaulting to the triple if present. Frankly, I find even the override case unconvincing.

Yes. If there is a target triple in the IR then overriding it might be weird. I don't remember exactly how it works in llc, but as a user I guess I'd probably would assume that the behavior would be the same between the tools when using cl option vs having the triple in the IR, or having both.

Nov 11 2021, 12:08 AM · Restricted Project

Nov 5 2021

markus committed rG7e34d5ead175: [NPM] Fix bug in llvm/utils/reduce_pipeline.py (authored by markus).
[NPM] Fix bug in llvm/utils/reduce_pipeline.py
Nov 5 2021, 12:42 AM
markus closed D113177: [NPM] Fix bug in llvm/utils/reduce_pipeline.py.
Nov 5 2021, 12:42 AM · Restricted Project

Nov 4 2021

markus requested review of D113177: [NPM] Fix bug in llvm/utils/reduce_pipeline.py.
Nov 4 2021, 3:04 AM · Restricted Project

Nov 2 2021

markus committed rGfd41738e2ca9: Recommit "[llvm-reduce] Add MIR support" (authored by markus).
Recommit "[llvm-reduce] Add MIR support"
Nov 2 2021, 2:17 AM
markus added a reverting change for rGbc2773cb1bdf: [llvm-reduce] Add MIR support: rGaee7f3384b40: Revert "[llvm-reduce] Add MIR support".
Nov 2 2021, 1:41 AM
markus committed rGaee7f3384b40: Revert "[llvm-reduce] Add MIR support" (authored by markus).
Revert "[llvm-reduce] Add MIR support"
Nov 2 2021, 1:41 AM
markus added a reverting change for D110527: [llvm-reduce] Add MIR support.: rGaee7f3384b40: Revert "[llvm-reduce] Add MIR support".
Nov 2 2021, 1:41 AM · Restricted Project
markus committed rGbc2773cb1bdf: [llvm-reduce] Add MIR support (authored by markus).
[llvm-reduce] Add MIR support
Nov 2 2021, 1:24 AM
markus closed D110527: [llvm-reduce] Add MIR support..
Nov 2 2021, 1:24 AM · Restricted Project

Oct 29 2021

markus added a comment to D110527: [llvm-reduce] Add MIR support..

Do these tools have anything in common other than the delta algorithm?

I would say that the tools are almost identical except that they work on different representations (IR and MIR). The purpose of the tools are the same, the command line interface, the handling of temporary files and the execution of interestingness tests are all the same and of course the main component the delta algorithm is the same.

Oct 29 2021, 12:47 AM · Restricted Project
markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Addressed review remarks.

Oct 29 2021, 12:21 AM · Restricted Project

Oct 28 2021

Herald added a reviewer for D110527: [llvm-reduce] Add MIR support.: luke957.

So how do we proceed?

Oct 28 2021, 6:18 AM · Restricted Project

Oct 22 2021

markus added a comment to D110527: [llvm-reduce] Add MIR support..

I'm not sure i fully like the coupling here.
In reality, how much code duplication is being avoided by bundling these two tools together?

Well, it is true that the llvm-reduce tool did not contain that many lines of code in the first place, and especially not if one excludes the IR specific reduction passes.
There is the Delta Debugging Algorithm in Delta.cpp and some general stuff for handling input and output files as well as running the interestingness tests.
In general code duplication is considered a bad thing so I think we all can agree that code reuse in some sense is desirable.

Would it not be better to extract the common interface and simplify writing new back/front -ends for llvm-reduce,
instead of having everything in a single tool?

Could be but it also seems like that would be a much larger change (at least to the existing llvm-reduce). I am not particularly interested in doing that in this patch but as always anyone is free to evolve things further in a follow up commit if they consider it worth while.

Oct 22 2021, 2:06 AM · Restricted Project
markus added inline comments to D110527: [llvm-reduce] Add MIR support..
Oct 22 2021, 1:41 AM · Restricted Project
markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Addressed review remarks.

Oct 22 2021, 1:36 AM · Restricted Project

Oct 20 2021

markus added a comment to D111531: Fix minor deficiency in MachineSink..

ping

Oct 20 2021, 4:56 AM · Restricted Project
markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Added -x flag to set input language as well as autodetection from filename.

Oct 20 2021, 4:34 AM · Restricted Project

Oct 15 2021

markus abandoned D55107: Prototype BasicAA improvements for discussion on llvm-dev.
Oct 15 2021, 6:17 AM

Oct 14 2021

markus added a comment to D111809: [DebugInfo] Correctly handle arrays with 0-width elements in GEP salvaging.

Can't say much besides that it makes sense to fix the problem at the source (and not appending to the list) instead of simply weakening the assert.

Oct 14 2021, 10:45 PM · Restricted Project, debug-info
markus added a comment to rGc72705678c47: Partial Reapply "[DebugInfo] Use variadic debug values to salvage BinOps and….

The reproducer is rather trivial as well but it did not immediately work outside our target so I had to fiddle a bit

opt -passes='function(dce)' bbi-61411-nophx.ll

should result in

opt: ../lib/Transforms/Utils/Local.cpp:1819: llvm::Value *getSalvageOpsForGEP(llvm::GetElementPtrInst *, const llvm::DataLayout &, uint64_t, SmallVectorImpl<uint64_t> &, SmallVectorImpl<llvm::Value *> &): Assertion `Offset.second.isStrictlyPositive() && "Expected strictly positive multiplier for offset."' failed.

Oct 14 2021, 6:26 AM
markus added inline comments to rGc72705678c47: Partial Reapply "[DebugInfo] Use variadic debug values to salvage BinOps and….
Oct 14 2021, 5:38 AM
markus added inline comments to rGc72705678c47: Partial Reapply "[DebugInfo] Use variadic debug values to salvage BinOps and….
Oct 14 2021, 4:59 AM
markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Renamed MyMachineModule to ReducerWorkItem
Got rid of raw function pointer.

Oct 14 2021, 2:29 AM · Restricted Project

Oct 13 2021

markus added inline comments to D110527: [llvm-reduce] Add MIR support..
Oct 13 2021, 8:17 AM · Restricted Project
markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Rebased to top of main and addressed some review remarks.

Oct 13 2021, 8:07 AM · Restricted Project

Oct 11 2021

markus added a reviewer for D110527: [llvm-reduce] Add MIR support.: diegotf.
Oct 11 2021, 4:08 AM · Restricted Project
markus requested review of D111531: Fix minor deficiency in MachineSink..
Oct 11 2021, 3:17 AM · Restricted Project

Oct 7 2021

markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Ok, so lets drop the [WIP] status and try to get this thing reviewed.

Oct 7 2021, 2:45 AM · Restricted Project

Oct 6 2021

markus committed rG91d15aa0b8bf: [NPM] Automatic 'opt' pipeline reducer script. (authored by markus).
[NPM] Automatic 'opt' pipeline reducer script.
Oct 6 2021, 12:11 AM
markus closed D110908: [NPM] Automatic 'opt' pipeline reducer script..
Oct 6 2021, 12:11 AM · Restricted Project

Oct 5 2021

markus added a comment to D110908: [NPM] Automatic 'opt' pipeline reducer script..

Thanks for reviewing!

Oct 5 2021, 11:25 PM · Restricted Project
markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Slightly better integration.

Oct 5 2021, 7:41 AM · Restricted Project
markus updated the diff for D110908: [NPM] Automatic 'opt' pipeline reducer script..

Added option for automatic expansion of default pipelines. E.g. now one can do

./utils/reduce_pipeline.py --opt-binary=../build-debug/bin/opt --input=bbi-60925.ll --output=test2.ll --passes='default<O3>' --expand-passes

and that will expand the pipeline (using a separate opt invocation with -print-pipeline-passes) before starting the reduction. In this particular case there will be no reduction as the crash does not reproduce with the default pipeline but that is a different matter. One test has been added for the expansion using fake_opt.py.

Oct 5 2021, 3:57 AM · Restricted Project
markus updated the diff for D110527: [llvm-reduce] Add MIR support..

Did a quick & dirty integration into llvm-reduce. All lit-tests as in llvm/test/tools/llvm-reduce/ pass but I suspect that the integration needs to be much prettier.

Oct 5 2021, 1:05 AM · Restricted Project

Oct 4 2021

markus added a comment to D110908: [NPM] Automatic 'opt' pipeline reducer script..

Great to see progress on this! For a bit of extra convenience, it might be helpful to accept -passes=defaults<O3> & co and automatically translate that to the explicit pass invocation.

Oct 4 2021, 5:36 AM · Restricted Project
markus updated the diff for D110908: [NPM] Automatic 'opt' pipeline reducer script..

Something like this and then perhaps add a few more tests and documenting them a bit?

Oct 4 2021, 4:15 AM · Restricted Project

Oct 1 2021

markus added a comment to D110908: [NPM] Automatic 'opt' pipeline reducer script..

Ok, cool. I guess it is sufficient to have the tests separated like that since they only need to be run when changes are made to the script.

Oct 1 2021, 9:57 AM · Restricted Project
markus added a comment to D110908: [NPM] Automatic 'opt' pipeline reducer script..

I thought about that too but where do we put tests for scripts in llvm/utils?

Oct 1 2021, 9:30 AM · Restricted Project
markus updated the diff for D110908: [NPM] Automatic 'opt' pipeline reducer script..

Added some docstrings.

Oct 1 2021, 5:38 AM · Restricted Project