Page MenuHomePhabricator

uabelho (Mikael Holmén)
User

Projects

User does not belong to any projects.

User Details

User Since
Dec 3 2015, 1:10 AM (251 w, 26 m)

Recent Activity

Today

uabelho committed rGa1217620a87f: [unittests] Use std::make_tuple to make some toolchains happy again (authored by uabelho).
[unittests] Use std::make_tuple to make some toolchains happy again
Thu, Sep 24, 2:26 AM

Fri, Sep 18

uabelho added a comment to D87149: [InstCombine] erase instructions leading up to unreachable.

We've run into a case where instcombine crashes with this patch:

opt -o /dev/null bbi-47401.ll -instcombine

on

@c = external global i16*, align 1
Fri, Sep 18, 12:32 AM · Restricted Project

Thu, Sep 17

uabelho committed rGbb037c2a7625: [ConstraintSystem] Remove local variable that is set but not read [NFC] (authored by uabelho).
[ConstraintSystem] Remove local variable that is set but not read [NFC]
Thu, Sep 17, 5:29 AM

Mon, Sep 14

uabelho added a comment to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

Hi,

I'm seeing what I think is a miscompile with this:

Thanks for the reproducer. We have to be more careful around pointers that may reference multiple locations in memory. I pushed f715d81c9df3 which limits elimination to stores that only reference a single location (or are in the same BB). The check is conservative and can be improved in the future, but the impact on the number of stores eliminated seems very low in practice. It would be great if you could check if that fixes the original reproducer.

Great! I'll check and get back during the day.

Mon, Sep 14, 5:52 AM · Restricted Project, Restricted Project
uabelho added a comment to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

Hi,

I'm seeing what I think is a miscompile with this:

Thanks for the reproducer. We have to be more careful around pointers that may reference multiple locations in memory. I pushed f715d81c9df3 which limits elimination to stores that only reference a single location (or are in the same BB). The check is conservative and can be improved in the future, but the impact on the number of stores eliminated seems very low in practice. It would be great if you could check if that fixes the original reproducer.

Mon, Sep 14, 4:50 AM · Restricted Project, Restricted Project
uabelho added a comment to D86689: [DAGCombine] Don't delete the node if it has uses immediately.

I wrote https://bugs.llvm.org/show_bug.cgi?id=47517 about a crash that starts happening with this patch.

Mon, Sep 14, 4:20 AM · Restricted Project
uabelho added a comment to D87163: [DSE] Switch to MemorySSA-backed DSE by default..

I'm seeing what I think is a miscompile with this:

opt -S -dse -o - dse.ll

on

Mon, Sep 14, 12:18 AM · Restricted Project, Restricted Project

Mon, Sep 7

uabelho committed rGea795304ec07: [PowerPC] Add parentheses to silence gcc warning (authored by uabelho).
[PowerPC] Add parentheses to silence gcc warning
Mon, Sep 7, 11:41 PM

Fri, Sep 4

uabelho added a comment to D86534: [GVN] Preserve MemorySSA if it is available..

The following crashes for me with this patch
opt -memoryssa -gvn -early-cse-memssa -S -o - foo.ll

Fri, Sep 4, 3:44 AM · Restricted Project

Thu, Sep 3

uabelho added inline comments to D86589: (Expensive) Check for Loop, SCC and Region pass return status.
Thu, Sep 3, 9:37 AM · Restricted Project
uabelho added inline comments to D86589: (Expensive) Check for Loop, SCC and Region pass return status.
Thu, Sep 3, 3:03 AM · Restricted Project

Wed, Sep 2

uabelho added a comment to D86839: [CMake] Remove -Wl,-allow-shlib-undefined which was added in rL221530.

We ran into problems with this commit.

Wed, Sep 2, 5:36 AM · Restricted Project
uabelho added inline comments to D86738: AMDGPU/GlobalISel: Implement computeKnownBits for groupstaticsize.
Wed, Sep 2, 2:42 AM · Restricted Project

Tue, Sep 1

uabelho added a comment to D85971: [IndVarSimplify] Fix Modified status for removal of overflow intrinsics.

I wrote
https://bugs.llvm.org/show_bug.cgi?id=47381
about a similar problem also caused by the missing setting of Changed in SimplifyIndvar::eliminateOverflowIntrinsic.

Tue, Sep 1, 4:21 AM · Restricted Project

Fri, Aug 28

uabelho added a comment to D71538: [SCEVExpander] Preserve LCSSA directly..

We ran into a problem with this patch. I wrote https://bugs.llvm.org/show_bug.cgi?id=47343 about it.

Fri, Aug 28, 4:12 AM · Restricted Project

Wed, Aug 26

uabelho added a comment to D86547: [compiler-rt][builtins] Use c[tl]zsi macro instead of __builtin_c[tl]z.

We've run with changes like this for a long time to make the routines work for our out of tree target where int is also 16 bits so thumbs up from me even if I'm not sure I can formally LGTM it.

Wed, Aug 26, 2:09 AM · Restricted Project

Tue, Aug 25

uabelho committed rG1c39ffecd84a: [libunwind] Fix warning when building without frameheader cache (authored by uabelho).
[libunwind] Fix warning when building without frameheader cache
Tue, Aug 25, 4:02 AM
uabelho committed rG59e1fbe55784: [PowerPC] Fix gcc warning [NFC] (authored by uabelho).
[PowerPC] Fix gcc warning [NFC]
Tue, Aug 25, 4:02 AM

Jul 16 2020

uabelho added inline comments to D83760: [AMDGPU][MC] Corrected MTBUF parsing and decoding.
Jul 16 2020, 10:56 PM · Restricted Project
uabelho committed rG274332282cb4: [clangd] Fix a few gcc warnings [NFC] (authored by uabelho).
[clangd] Fix a few gcc warnings [NFC]
Jul 16 2020, 1:16 AM
uabelho committed rGae74387fc0fd: [MasmParser] Remove unused method emitStructValue to silence warning (authored by uabelho).
[MasmParser] Remove unused method emitStructValue to silence warning
Jul 16 2020, 1:16 AM

Jul 10 2020

uabelho added a comment to D82998: [BasicAA] Enable -basic-aa-recphi by default.

Thanks. I will take a look.

From this, that bug appears to be fixed: https://godbolt.org/z/6W8hc8. Probably from D82987 I would expect. I will check. Opt 10.0.0 does show the problem, where as trunk seems to be fixed.

Let me know about the runtime failure.

Jul 10 2020, 3:00 AM · Restricted Project
uabelho added a comment to D82998: [BasicAA] Enable -basic-aa-recphi by default.

I noticed that we got a runtime failure with this patch for our out-of-tree target. I haven't gone to the bottom with it yet, but I saw that a colleague of mine wrote a PR about a miscompile with -basic-aa-recphi that at least hasn't been handled according to bugzilla:
https://bugs.llvm.org/show_bug.cgi?id=37952

Jul 10 2020, 1:32 AM · Restricted Project

Jul 9 2020

uabelho added a comment to D81345: [LV] Vectorize without versioning-for-unit-stride under -Os/-Oz.

I start seeing a crash with this patch:
https://bugs.llvm.org/show_bug.cgi?id=46652

Jul 9 2020, 1:11 AM · Restricted Project

Jul 8 2020

uabelho added inline comments to D75306: [ms] [llvm-ml] Add initial MASM STRUCT/UNION support.
Jul 8 2020, 1:58 AM · Restricted Project
uabelho committed rGcfcf8e17ef53: [analyzer] Silence gcc -Wparentheses warning [NFC] (authored by uabelho).
[analyzer] Silence gcc -Wparentheses warning [NFC]
Jul 8 2020, 1:26 AM

Jul 5 2020

uabelho added a comment to D82490: [libcxx] Put clang::trivial_abi on std::unique_ptr.

A whole bunch of buildbots seems to have failed as well (in different ways).
E.g.
http://lab.llvm.org:8011/builders/llvm-clang-win-x-armv7l/builds/227/steps/test-libc%2B%2B/logs/FAIL%3A%20libc%2B%2B%3A%3Aunique_ptr_arg.pass.cpp
hit the same assertion as I see.
How about reverting again?

Jul 5 2020, 11:57 PM · Restricted Project
uabelho added a comment to D82490: [libcxx] Put clang::trivial_abi on std::unique_ptr.

I hit this assertion in a couple of testcases when running check-all on trunk now:

Jul 5 2020, 11:27 PM · Restricted Project

Jun 30 2020

uabelho added a comment to D82453: [CostModel] Fix for cast crash.

I don't know this code but at least the crash I saw goes away with the fix. Thanks!

Jun 30 2020, 2:08 AM · Restricted Project

Jun 24 2020

uabelho added a comment to D81643: [NFCI][CostModel] Unify ExtractElement cost..

I wrote
https://bugs.llvm.org/show_bug.cgi?id=46430
about a crash that started happening with this patch.

Jun 24 2020, 1:34 AM · Restricted Project

Jun 23 2020

uabelho added a comment to D81920: [clangd] Change FSProvider::getFileSystem to take CurrentWorkingDirectory.

Can you check if you see the warning with different versions of gcc?

Jun 23 2020, 11:57 PM · Restricted Project
uabelho added a comment to D82029: [Coroutines] Ensure co_await promise.final_suspend() does not throw.

Test failures are being fixed in https://reviews.llvm.org/D82338/new/

Jun 23 2020, 10:08 PM · Restricted Project
uabelho committed rG82a882db0832: Silence gcc warning about extra ";" [NFC] (authored by uabelho).
Silence gcc warning about extra ";" [NFC]
Jun 23 2020, 2:36 AM
uabelho added inline comments to D81920: [clangd] Change FSProvider::getFileSystem to take CurrentWorkingDirectory.
Jun 23 2020, 2:36 AM · Restricted Project
uabelho added a comment to D82029: [Coroutines] Ensure co_await promise.final_suspend() does not throw.

I see a bunch of failures when I run libcxx testcases with this patch. Should there be "noexcept" on a number of more places?

Jun 23 2020, 2:04 AM · Restricted Project
uabelho added inline comments to D81920: [clangd] Change FSProvider::getFileSystem to take CurrentWorkingDirectory.
Jun 23 2020, 1:00 AM · Restricted Project

Jun 17 2020

uabelho added inline comments to D81860: [ARM] Fix crash trying to generate i1 immediates.
Jun 17 2020, 2:07 AM · Restricted Project

Jun 16 2020

uabelho added a comment to D80206: [IR] Clean up dead instructions after simplifying a conditional branch.

I wrote
https://bugs.llvm.org/show_bug.cgi?id=46343
about a crash that starts happening with this patch.

Jun 16 2020, 2:09 AM · Restricted Project

Jun 11 2020

uabelho added a comment to D81634: Remove assert from ShuffleVectorInst.

Was the failure only a fuzzer problem?

Jun 11 2020, 10:29 PM · Restricted Project
uabelho added a comment to D79218: Process gep (phi ptr1, ptr2) in SROA.

Hi!

I wrote
https://bugs.llvm.org/show_bug.cgi?id=46280
for a crash that started happening with this commit.

Thanks! Here is the fix: D81674

Jun 11 2020, 10:29 PM · Restricted Project
uabelho added a comment to D79218: Process gep (phi ptr1, ptr2) in SROA.

I wrote
https://bugs.llvm.org/show_bug.cgi?id=46280
for a crash that started happening with this commit.

Jun 11 2020, 6:35 AM · Restricted Project
uabelho resigned from D81634: Remove assert from ShuffleVectorInst.

Thanks for the patch so quickly!
I really have no idea if the assert actually makes sense or not though, so it's best if @spatel can have an opinion.

Jun 11 2020, 1:50 AM · Restricted Project
uabelho added a comment to D80992: [CostModel] Unify getArithmeticInstrCost .

I wrote a PR about a crash that starts occuring with this patch:
https://bugs.llvm.org/show_bug.cgi?id=46276

Jun 11 2020, 12:37 AM · Restricted Project

Jun 10 2020

uabelho added a comment to D74825: SpeculativeExecution: fixed ingoring free execution.

I just wrote a PR about a debug-info fault that starts occuring with this patch:
https://bugs.llvm.org/show_bug.cgi?id=46267

Jun 10 2020, 2:08 AM · Restricted Project

Jun 9 2020

uabelho added a comment to D81408: [builtins] Improve compatibility with 16 bit targets.

Our out-of-tree target also have 16 bit ints and we've had to change a whole bunch of builtin_clz to clzsi and builtin_ctz to ctzsi to make things work so this patch is a step in the right direction for us too.

Jun 9 2020, 7:05 AM · Restricted Project

Jun 4 2020

uabelho added a comment to D71943: AMDGPU/GlobalISel: Add stub reg-bank aware combiner pass.

With gcc 7.4, the following warning pops up with this patch:

Jun 4 2020, 7:04 AM · Restricted Project
uabelho added inline comments to D77644: [clangd] Handle additional includes while parsing ASTs.
Jun 4 2020, 2:40 AM · Restricted Project
uabelho committed rG2f671c42254d: [WebAssembly] Fix gcc warning [NFC] (authored by uabelho).
[WebAssembly] Fix gcc warning [NFC]
Jun 4 2020, 1:35 AM

Jun 3 2020

uabelho added inline comments to D54368: RegAllocFast: Record internal state based on register units.
Jun 3 2020, 11:57 PM · Restricted Project
uabelho added a comment to D80863: [WebAssembly] Eliminate range checks on br_tables.
In D80863#2070105, @kcc wrote:

Hi, 
This made our ubsan bots red. Please fix or revert ASAP
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/42256

Jun 3 2020, 1:03 AM · Restricted Project

May 27 2020

uabelho added a comment to D78987: [BPI][NFC] Reuse post dominantor tree from analysis manager when available.

Hi!
PR 46098 started occuring wiith this patch:
https://bugs.llvm.org/show_bug.cgi?id=46098

May 27 2020, 6:27 AM · Restricted Project

May 19 2020

uabelho committed rGe408b54290bd: [clangd] Fix gcc compiler warning by removing extra ";" [NFC] (authored by uabelho).
[clangd] Fix gcc compiler warning by removing extra ";" [NFC]
May 19 2020, 3:13 AM

May 14 2020

uabelho added a comment to D79467: [PDB] Optimize public symbol processing.
In D79467#2035176, @rnk wrote:

Hopefully rG409274274 addresses the GCC warnings. I didn't get a chance to locally repro and confirm that the fix works.

May 14 2020, 1:01 AM · Restricted Project

May 13 2020

uabelho added inline comments to D79467: [PDB] Optimize public symbol processing.
May 13 2020, 3:11 AM · Restricted Project
uabelho added inline comments to D79467: [PDB] Optimize public symbol processing.
May 13 2020, 2:39 AM · Restricted Project

May 12 2020

uabelho added a comment to D76483: [DivRemPairs] Freeze operands if they can be undef values.

I wrote https://bugs.llvm.org/show_bug.cgi?id=45885 about a crash which starts happening with this patch.

May 12 2020, 4:47 AM · Restricted Project

May 7 2020

uabelho added a comment to D79510: [PATCH] When pragma FENV_ACCESS is ignored do not modify Sema.CurFPFeatures.

Great! Thanks!

May 7 2020, 12:33 AM · Restricted Project

May 6 2020

uabelho added a comment to D72841: Add support for pragma float_control, to control precision and exception behavior at the source level.

It seems to be this change in SemaStmt.cpp that makes the FENV-pragma have some effect regardless of the warning saying that the pragma doesn't have any effect:

May 6 2020, 3:44 AM · Restricted Project, Restricted Project
uabelho added a comment to D72841: Add support for pragma float_control, to control precision and exception behavior at the source level.

I think I'm seeing some oddities with this patch (current trunk, 0054c46095).
With

May 6 2020, 3:11 AM · Restricted Project, Restricted Project

Apr 29 2020

uabelho added inline comments to D78662: [builtins] Support architectures with 16-bit int.
Apr 29 2020, 5:51 AM · Restricted Project

Apr 16 2020

uabelho added a comment to D77620: [SimpleLoopUnswitch] Do not delete DT edge when a duplicate exists..

While this solution does solve the problem, there's an underlying bug that this is hiding. I've talking with Chandler about this and I'll send a couple of patches that are aiming to fix the underlying problem instead.
Apologies for the delay in getting you a fix.

Apr 16 2020, 12:54 AM · Restricted Project

Apr 15 2020

uabelho accepted D77620: [SimpleLoopUnswitch] Do not delete DT edge when a duplicate exists..

I've run quite a number of tests with this without problems.

Apr 15 2020, 11:25 PM · Restricted Project

Apr 14 2020

uabelho added a comment to D77539: [LV] Add VPValue operands to VPBlendRecipe (NFCI).

I wrote
https://bugs.llvm.org/show_bug.cgi?id=45525
for a crash that started appearing with this patch.

Apr 14 2020, 1:33 AM · Restricted Project

Apr 8 2020

uabelho added a comment to D77651: [IfConversion] Disallow TrueBB == FalseBB for valid diamonds.

Thanks!
Submitted in 893df2032d4.

Apr 8 2020, 4:18 AM · Restricted Project
uabelho committed rG893df2032d48: [IfConversion] Disallow TrueBB == FalseBB for valid diamonds (authored by uabelho).
[IfConversion] Disallow TrueBB == FalseBB for valid diamonds
Apr 8 2020, 4:18 AM
uabelho closed D77651: [IfConversion] Disallow TrueBB == FalseBB for valid diamonds.
Apr 8 2020, 4:17 AM · Restricted Project

Apr 7 2020

uabelho created D77651: [IfConversion] Disallow TrueBB == FalseBB for valid diamonds.
Apr 7 2020, 8:39 AM · Restricted Project
uabelho added a comment to D77651: [IfConversion] Disallow TrueBB == FalseBB for valid diamonds.

I have no idea if this is the best fix for
https://bugs.llvm.org/show_bug.cgi?id=45302
but I saw that we've made a similar fix for triangles in 957e090ac9 and thought maybe this could be an easy way out in this case.

Apr 7 2020, 8:39 AM · Restricted Project
uabelho updated the summary of D77651: [IfConversion] Disallow TrueBB == FalseBB for valid diamonds.
Apr 7 2020, 8:39 AM · Restricted Project
uabelho added a comment to D77620: [SimpleLoopUnswitch] Do not delete DT edge when a duplicate exists..

I don't really know this code but at least the fix avoids the crash I saw. Thanks!

Apr 7 2020, 2:40 AM · Restricted Project

Mar 31 2020

uabelho added a comment to D77160: Fix PR45371: SeparateConstOffsetFromGEP clean up bookkeeping.

Thanks! This indeed seem to solve the problem I saw.

Mar 31 2020, 11:58 PM · Restricted Project

Mar 27 2020

uabelho committed rG7d482e921353: Fix TBAA for unsigned fixed-point types (authored by uabelho).
Fix TBAA for unsigned fixed-point types
Mar 27 2020, 2:39 AM
uabelho closed D76856: Fix TBAA for unsigned fixed-point types.
Mar 27 2020, 2:38 AM · Restricted Project

Mar 13 2020

uabelho committed rGabc238a3c736: [libunwind] Silence warnings when __mips_hard_float is not defined (authored by uabelho).
[libunwind] Silence warnings when __mips_hard_float is not defined
Mar 13 2020, 1:33 AM

Mar 5 2020

uabelho committed rGcd1dc7f15d63: [AST] Replace assert with llvm_unreachable to silence compiler warning (authored by uabelho).
[AST] Replace assert with llvm_unreachable to silence compiler warning
Mar 5 2020, 2:46 AM

Feb 24 2020

uabelho added a comment to D74792: [SimplifyLibCalls][IRBuilder] Accept any IRBuilder in SimplifyLibCalls.

@nikic Alright, thanks!

Feb 24 2020, 1:45 AM · Restricted Project
uabelho added a comment to D74792: [SimplifyLibCalls][IRBuilder] Accept any IRBuilder in SimplifyLibCalls.

Is this supposed to affect the output from instcombine?

Feb 24 2020, 12:42 AM · Restricted Project

Feb 21 2020

uabelho added a comment to D74901: [ms] [llvm-ml] Remove unused functions.

I have no idea if it's reasonable to remove these methods but it does indeed silence the compiler warnings about unused functions that I saw. Thanks for that.

Feb 21 2020, 12:14 AM · Restricted Project

Feb 10 2020

uabelho committed rGa50c0b0df733: Fix compiler warning when compiling without asserts [NFC] (authored by uabelho).
Fix compiler warning when compiling without asserts [NFC]
Feb 10 2020, 5:03 AM

Jan 27 2020

uabelho added inline comments to D51028: [BranchFolder] Drop kill/dead flags if they aren't present in all merged instructions.
Jan 27 2020, 2:10 AM · Restricted Project
uabelho added a comment to D51028: [BranchFolder] Drop kill/dead flags if they aren't present in all merged instructions.

I am a little curious as to how these differences in flags arise. I can understand that a kill flag can be seen as an optimization and is optional. The same would go for a dead flag. But I am not sure about an undef flag... If it is present in one case but not in the other, the two instructions do not seem equivalent, or?

So I suppose your patch makes sense to me, although I am not familiar enough to understand the undef case... Perhaps some comment on how these three cases can emerge and may be treated would be good? (at least to me :-)

Jan 27 2020, 1:37 AM · Restricted Project

Jan 16 2020

uabelho added a comment to D72734: [VectorUtils] Rework the Vector Function Database (VFDatabase)..

If you are happy with it, may I ask you to formally approve it so I can submit it? (Feel free to do a full review if you prefer to do so!)

Jan 16 2020, 2:54 AM · Restricted Project

Jan 15 2020

uabelho added a comment to D72709: [codegen,amdgpu] Enhance MIR DIE and re-arrange it for AMDGPU..

In my out-of-tree target we are running DeadMachineInstructionElim after SSA, and now of course the new assert fails.

Jan 15 2020, 11:39 PM · Restricted Project

Jan 14 2020

uabelho added a comment to D72734: [VectorUtils] Rework the Vector Function Database (VFDatabase)..

@uabelho , let me know if the introduction of isTLIScalarize prevents the regressions you have seen in your out-of-tree compiler.

Jan 14 2020, 11:58 PM · Restricted Project

Jan 13 2020

uabelho added a comment to D67572: [VectorUtils] Introduce the Vector Function Database (VFDatabase)..

Would that work for you?

Jan 13 2020, 10:43 PM · Restricted Project

Jan 8 2020

uabelho committed rG8e2b44f7e064: [Intrinsic] Add fixed point division intrinsics. (authored by ebevhan).
[Intrinsic] Add fixed point division intrinsics.
Jan 8 2020, 6:26 AM
uabelho closed D70007: [Intrinsic] Add fixed point division intrinsics..
Jan 8 2020, 6:25 AM · Restricted Project
uabelho added a comment to D70007: [Intrinsic] Add fixed point division intrinsics..

I'll submit this for @ebevhan in a little bit.

Jan 8 2020, 6:16 AM · Restricted Project

Dec 19 2019

uabelho removed a watcher for debug-info: uabelho.
Dec 19 2019, 4:59 AM
uabelho added reviewers for D51028: [BranchFolder] Drop kill/dead flags if they aren't present in all merged instructions: qcolombet, MatzeB, jonpa.
Dec 19 2019, 12:49 AM · Restricted Project

Dec 17 2019

uabelho updated the diff for D51028: [BranchFolder] Drop kill/dead flags if they aren't present in all merged instructions.

Rebased, updated to drop dead flags as well.

Dec 17 2019, 6:01 AM · Restricted Project

Dec 13 2019

uabelho added a comment to D67572: [VectorUtils] Introduce the Vector Function Database (VFDatabase)..

Ah, OK. I see what you mean. I think we could solve this by scalarizing after vectorization, instead of doing it in the vectorizer.

We could have a IR pass that runs after the vectorizer that looks for intrinsics calls .When it sees an intrinsics that operates on vector, it checks whether the target is able to lower it to some function or instruction. If not, it scalarizes it. With this we wouldn't have to introduce special behavior in the vectorizer for handling intrinsics: it could just vectorize any call to intrinsics for which vectorization make sense.

Dec 13 2019, 3:12 AM · Restricted Project
uabelho updated subscribers of D67572: [VectorUtils] Introduce the Vector Function Database (VFDatabase)..
Dec 13 2019, 12:16 AM · Restricted Project

Dec 12 2019

uabelho added a comment to D67572: [VectorUtils] Introduce the Vector Function Database (VFDatabase)..

@uabelho: would it be possible for you to provide me a minimal reproducer that I could use to craft the (wip) solution I have in mind?

Dec 12 2019, 11:28 PM · Restricted Project
uabelho added a comment to D67572: [VectorUtils] Introduce the Vector Function Database (VFDatabase)..

[...]

Is this change in behavior intended?

This change in behavior sounds a bit worrying for down-stream targets.

I think we should have at least an assertion making sure the check in LoopVEctorizationLegality succeeds in all cases it did previously with isFunctionVectorizable, otherwise down-stream targets will silently miss out on vectorization.

Dec 12 2019, 2:54 AM · Restricted Project
uabelho added a comment to D67572: [VectorUtils] Introduce the Vector Function Database (VFDatabase)..

Thanks for the reply!

Dec 12 2019, 1:57 AM · Restricted Project

Dec 11 2019

uabelho added a comment to D67572: [VectorUtils] Introduce the Vector Function Database (VFDatabase)..

A question regarding this patch.
For my out-of-tree target vectorization of intrinsics added for my target seems to have stopped working with this patch.
Is there something/what do I have to do to make the vectorizer understand my intrinsics are vectorizable?

Dec 11 2019, 6:54 AM · Restricted Project

Dec 10 2019

uabelho added a comment to D71279: [DebugInfo][MachineSink] Don't stop copy-propagating DBG_VALUEs that will have a sunk copy created.

I think you're on the right track. My mir-reproducer passes with this patch, so now I get the wanted

Dec 10 2019, 11:19 PM · Restricted Project
uabelho added a comment to D70676: [DebugInfo] Don't repeatedly created undef DBG_VALUEs during machine-sinking.

D71283 Should (TM) fix it, but, the x86 test that I wrote and which seemed to replicate the issue, now no longer replicates the issue. Which is unfortunate.

Would you be able to give that patch (and its parent) a test to see whether it fixes the problem you're encountering? Storing the subregister separately should entirely fix this I believe, I'm just having problems replicating it on x86.

Dec 10 2019, 11:19 PM · Restricted Project
uabelho added a comment to D71283: [WIP][DebugInfo][MachineSink] Preserve DBG_VALUE subregisters when sinking.

I think you're on the right track. My mir-reproducer passes with this patch, so now I get the wanted

Dec 10 2019, 11:10 PM · Restricted Project