dtemirbulatov (Dinar Temirbulatov)
User

Projects

User does not belong to any projects.

User Details

User Since
Sep 17 2015, 10:06 AM (148 w, 3 d)

Recent Activity

Tue, Jul 10

dtemirbulatov accepted D49135: [SLPVectorizer] Add initial alternate opcode support for cast instructions..

looks good for me.

Tue, Jul 10, 8:22 PM

Thu, Jul 5

dtemirbulatov added a comment to D48945: [SLPVectorizer] Begin abstracting InstructionsState alternate matching away from opcodes. NFCI..

Looks good for me.

Thu, Jul 5, 3:03 AM

Mon, Jul 2

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Mon, Jul 2, 12:22 PM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Mon, Jul 2, 12:17 PM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

here is change before 74cd05c4a4f94a27daf2d3fc173e7213060cc47c commit, I am currently rebaseing the change.

Mon, Jul 2, 12:15 PM
dtemirbulatov added a comment to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Hi, when do you plan to update the path for the latest version? I have an idea of improving the scheduling model, maybe it will help us to resolve the problems with the patch.

Mon, Jul 2, 9:38 AM

Sat, Jun 30

dtemirbulatov added a comment to D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

@dtemirbulatov What is the status of this patch please?

@RKSimon I think it needs review. I will rebase proposed change.

Sat, Jun 30, 8:06 AM

Jun 20 2018

dtemirbulatov added a comment to D48359: [SLPVectorizer] Use InstructionsState to record AltOpcode.

looks good for me.

Jun 20 2018, 4:56 AM

Jun 13 2018

dtemirbulatov added a comment to D48120: [SLPVectorizer] Remove RawInstructionsData/getMainOpcode and merge into getSameOpcode.

looks good for me.

Jun 13 2018, 11:17 AM

Jun 3 2018

dtemirbulatov added a comment to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

@dtemirbulatov @ABataev Is there anything that can be done to please keep this patch moving? The next release branch won't be far away now and I originally requested this over a year and a half ago (PR30787), and ideally I'd like to get this in and then I can more easily implement PR33744 as well in time.

I think that algorithm at 4512 is incorrect, we need to copy common instructions for bundles, instead of reordering bundles. Alexey?

Jun 3 2018, 9:35 AM

May 22 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
May 22 2018, 6:34 PM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
May 22 2018, 6:29 PM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
May 22 2018, 6:00 PM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Implement post scheduling bundle reorder of the last element of the bundle according how it was scheduled.

May 22 2018, 12:19 AM

Apr 23 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Apr 23 2018, 4:09 AM

Apr 22 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Apr 22 2018, 6:21 AM

Apr 16 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Apr 16 2018, 2:29 PM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update after remarks from Alexey Bataev.

Apr 16 2018, 2:21 PM

Apr 11 2018

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Rebased, improved complexity of setInsertPointAfterBundle() for already scheduled instructions, minor changes in scheduleBlock()

Apr 11 2018, 1:02 PM

Mar 31 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Mar 31 2018, 8:11 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update after Alexey's remarks.

Mar 31 2018, 7:54 AM

Mar 29 2018

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update after tests commit, formatting, delete Instruction::ExtractElement restricktion from tryToRepresentAsInstArg.

Mar 29 2018, 4:48 PM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Removed internal vector dependency check as incorrect and I added the dependency check in case of partial bundle vectorization with non-alternative operations only. At BoUpSLP::vectorizeTree, BoUpSLP::buildTree.

Mar 29 2018, 4:38 AM

Feb 13 2018

dtemirbulatov added a comment to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

I will update the solution shortly, I am currently in changing scheduler in order to avoid non-alternative opcodes in a bundle.

Feb 13 2018, 7:11 AM

Jan 30 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Jan 30 2018, 7:52 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Jan 30 2018, 7:32 AM

Jan 29 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Jan 29 2018, 1:39 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update after ABataev remarks.

Jan 29 2018, 1:29 AM

Jan 28 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Jan 28 2018, 12:25 PM

Jan 26 2018

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Remove dominance usage from the scheduler by adding support for nonalternative opcodes in calculateDependencies.

Jan 26 2018, 4:41 AM

Jan 25 2018

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Jan 25 2018, 4:03 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Jan 25 2018, 3:51 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Jan 25 2018, 3:42 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Fixed more scheduling errors, still scheduling algorithm could be improved in term of complexity, but I tried to minimally impact existing scheduling.

Jan 25 2018, 3:41 AM

Jan 11 2018

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update proposed change, fixed scheduling issue by avoiding using dominators.

Jan 11 2018, 9:25 PM

Dec 20 2017

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 20 2017, 7:20 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 20 2017, 2:19 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 20 2017, 1:40 AM

Dec 18 2017

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 18 2017, 8:36 PM

Dec 16 2017

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 16 2017, 7:13 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 16 2017, 7:10 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 16 2017, 7:07 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Rebase, fixed issue with the load instruction, add inner vector dependency check.

Dec 16 2017, 7:00 AM

Dec 13 2017

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 13 2017, 7:47 AM
dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Dec 13 2017, 2:49 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Fix all issues related to the last commit attempt.

Dec 13 2017, 2:46 AM

Nov 28 2017

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Nov 28 2017, 10:53 PM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Fix all issues related to the last commit attempt.

Nov 28 2017, 10:44 PM

Nov 22 2017

dtemirbulatov updated the diff for D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Update description in LangRef.rst.

Nov 22 2017, 4:10 PM
dtemirbulatov updated the diff for D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Update description in LangRef.rst.

Nov 22 2017, 4:09 PM

Nov 21 2017

dtemirbulatov updated the diff for D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Rebase and fix remarks.

Nov 21 2017, 2:44 PM
dtemirbulatov updated the diff for D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Rebase and fix remarks.

Nov 21 2017, 2:42 PM

Nov 13 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update after NFC with add test/Transforms/SLPVectorizer/SystemZ/lit.local.cfg to the tree.

Nov 13 2017, 10:39 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Fixed two issues associated with last build failures.

  1. llvm IR standard violation with second parameter always as integer 32-bit for Shl, LShr, AShr, but according to https://llvm.org/docs/LangRef.html#lshr-instruction both arguments of those instructions must be the same type. I have a testcase for the issue, but I think it is not required, please comment if you think otherwise.
  2. allow SystemZ SLP tests only when SystemZ is supported in a build.
Nov 13 2017, 10:16 AM

Nov 7 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Formatted.

Nov 7 2017, 11:43 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Fixed typo.

Nov 7 2017, 11:32 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update to the criteria to "(SameOrAlt <= VL.size() / 2) return;". I already tested that solution before.

Nov 7 2017, 11:28 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Updated this criteria for the same or alternative opcodes in the way suggested by Alexey and I tested the solution widely on different platforms and I was not able to encounter code where the same or alternative opcodes are more than 1 or 2 or the same or alternative opcode > than operations with different opcodes, except equal to one. But still the new criteria looks better to me.

Nov 7 2017, 10:43 AM

Oct 25 2017

dtemirbulatov updated the diff for D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Add LangRef.rst change. Small changes in SLPVectorizer.cpp.

Oct 25 2017, 3:38 PM
dtemirbulatov updated the diff for D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Add LangRef.rst change, update test to check metadata values.

Oct 25 2017, 3:35 PM
dtemirbulatov updated the diff for D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Update for solution, here it uses metadata "speculation.marker" with set of offsets instead of maximum offset as before, thread for discussion http://lists.llvm.org/pipermail/llvm-dev/2017-September/117782.html.

Oct 25 2017, 12:26 AM
dtemirbulatov updated the diff for D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Update for solution, here it uses metadata "speculation.marker" with set of offsets instead of maximum offset as before, thread for discussion http://lists.llvm.org/pipermail/llvm-dev/2017-September/117782.html.

Oct 25 2017, 12:25 AM

Oct 10 2017

dtemirbulatov added a comment to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Ping

Oct 10 2017, 5:44 AM

Oct 9 2017

dtemirbulatov added inline comments to D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..
Oct 9 2017, 6:06 AM

Oct 5 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Found trash in the previous change, Update after fix ABataev remarks.

Oct 5 2017, 2:17 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update after ABataev remarks

Oct 5 2017, 2:10 AM

Oct 3 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update with solution for the last commit issues with that looks more appropriate to me with just avoiding any vectors like [ operand, operand, operation, operand] or [load, load, operation, load] in a scheduling bundle, with this solution I have not noticed any bitcode regressions(efficiency) with provided testcase pr34619.ll. I have checked all list of issues from the previous build with this change.

Oct 3 2017, 3:17 AM

Oct 1 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Resolved all issues associated with previous commit.

Oct 1 2017, 5:34 AM

Sep 26 2017

dtemirbulatov added a comment to D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

I think something might be missing here. You're forming a 4x wide load, but you've only proven dereferenceability for offsets 0, 1, 3. (i.e. not 2). How do we know it's safe to dereference between the two elements 1 & 3?

well, it is a maximum reference that was recorded there in InstCombine, that implies 0, 1, 2 and 3. And if you are saying that 0, 1, 3 dereferenceable and 2-nd is not at the same time then this solution is not correct.

Sep 26 2017, 3:48 PM
dtemirbulatov added a comment to D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

are you still looking at some other kind of metadata solution? Did you mean for "metadata" here to mean the intrinsic you're proposing?

No, I am happy with intrinsic solution.

Sep 26 2017, 3:30 PM

Sep 25 2017

dtemirbulatov updated the diff for D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Fix review remarks and add an aggregate pointer as intrinsic's second parameter, remove getNextNode() usage.

Sep 25 2017, 5:02 AM
dtemirbulatov updated the diff for D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Fix recordDeadLoad() parameter lining with clang-format.

Sep 25 2017, 5:02 AM
dtemirbulatov updated the diff for D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Fix review remarks and add an aggregate pointer intrinsic's second parameter.

Sep 25 2017, 5:02 AM
dtemirbulatov updated the diff for D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

add missed test in the last change.

Sep 25 2017, 5:02 AM
dtemirbulatov added a comment to D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Thanks for working on this. Please post an RFC on llvm-dev about the intrinsic and how you intend to use it. You can reference this review, but we should make sure that we have design consensus before proceeding here.

Sep 25 2017, 5:02 AM
dtemirbulatov added a comment to D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Why not leave the original load in place, and tag the load metadata like !speculation_marker ? In either case, this needs a LangRef entry.

Re-reading, I messed up the phrasing there. I meant to say two things:

  • You could consider leaving the original load in place, and tag it with something like !speculation_marker. This metadata would indicate to passes like DCE that said load should not be eliminated till the very end of the optimization pipeline (so that we keep the optimization information for as long as possible). This does block optimizations though (like you will not be able to sink stores across this now present load, something you may have been able to do with the intrinsic you're introducing, though not without figuring out a way to denote it as not reading memory), but it has the upside of not introducing a new intrinsic.
  • Irrespective of whether you decide to go with the intrinsic or the metadata approach, please also update the language reference with an entry describing the semantics of this new thing.
Sep 25 2017, 5:02 AM
dtemirbulatov added a comment to D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.

Some style comments as the poor bracing/indentation makes it difficult to grok - more thorough review needs to wait until you've addressed @hfinkel's comments on D37648 about phantom mem intrinsics

Sep 25 2017, 5:02 AM

Sep 14 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Update after RKSimon's remarks.

Sep 14 2017, 2:57 AM

Sep 11 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Found typo in previous change. Update.

Sep 11 2017, 3:45 AM
dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Rebase

Sep 11 2017, 2:31 AM

Sep 8 2017

dtemirbulatov created D37648: [SLPVectorizer] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.
Sep 8 2017, 4:08 PM

Sep 7 2017

dtemirbulatov created D37579: [InstCombine] Fix PR21780 Expansion of 256 bit vector loads fails to fold into shuffles.
Sep 7 2017, 11:29 AM

Sep 6 2017

dtemirbulatov updated the diff for D37212: [SLPVectorizer] Add struct InstructionsState, NFC.

reduce the change a bit further and rebase.

Sep 6 2017, 8:15 AM

Sep 4 2017

dtemirbulatov added a comment to D37212: [SLPVectorizer] Add struct InstructionsState, NFC.

ping

Sep 4 2017, 6:13 AM

Aug 31 2017

dtemirbulatov updated the diff for D37212: [SLPVectorizer] Add struct InstructionsState, NFC.

rebase

Aug 31 2017, 3:05 PM

Aug 28 2017

dtemirbulatov created D37212: [SLPVectorizer] Add struct InstructionsState, NFC.
Aug 28 2017, 6:23 AM

Aug 20 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Rebase

Aug 20 2017, 10:45 PM

Aug 18 2017

dtemirbulatov updated the diff for D36766: Reorder operands with provided Opcode.

Remove getDefaultConstantForOpcode function for now.

Aug 18 2017, 5:49 AM

Aug 17 2017

dtemirbulatov updated the diff for D36766: Reorder operands with provided Opcode.

test update to simpler

Aug 17 2017, 7:01 PM

Aug 16 2017

dtemirbulatov updated the diff for D36766: Reorder operands with provided Opcode.

remove extra flag -slp-vectorizer in test.

Aug 16 2017, 7:57 PM
dtemirbulatov updated the diff for D36766: Reorder operands with provided Opcode.

update testcase.

Aug 16 2017, 7:53 PM

Aug 15 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

update after RKSimon's remarks.

Aug 15 2017, 9:08 PM
dtemirbulatov created D36766: Reorder operands with provided Opcode.
Aug 15 2017, 1:51 PM

Aug 14 2017

dtemirbulatov updated the diff for D28907: [SLP] Fix for PR30787: Failure to beneficially vectorize 'copyable' elements in integer binary ops..

Rebase

Aug 14 2017, 7:45 PM
dtemirbulatov updated the diff for D36518: [SLPVectorizer] Schedule bundle with different opcodes..

correct test filename spelling.

Aug 14 2017, 2:40 AM

Aug 13 2017

dtemirbulatov updated the diff for D36518: [SLPVectorizer] Schedule bundle with different opcodes..

update test.

Aug 13 2017, 9:53 PM
dtemirbulatov added inline comments to D36518: [SLPVectorizer] Schedule bundle with different opcodes..
Aug 13 2017, 9:40 PM

Aug 12 2017

dtemirbulatov updated the diff for D36518: [SLPVectorizer] Schedule bundle with different opcodes..

Update after RKSimon's remarks and replace testcase for more appropriate.

Aug 12 2017, 12:09 PM

Aug 10 2017

dtemirbulatov updated the diff for D36518: [SLPVectorizer] Schedule bundle with different opcodes..

update after RKSimon remarks and add test.

Aug 10 2017, 8:05 PM
dtemirbulatov updated the diff for D36518: [SLPVectorizer] Schedule bundle with different opcodes..

Test cases?

Well, I could add tests here, but it is a like a main flow of this algorithm and it should be tested already here.

Aug 10 2017, 4:55 AM