Page MenuHomePhabricator

CarolineConcatto (Caroline)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 4 2020, 3:58 AM (126 w, 1 d)

Recent Activity

May 20 2022

CarolineConcatto committed rG4cd951042682: [AArch64][SME]Tied up ZA operand for accumulate instructions (authored by CarolineConcatto).
[AArch64][SME]Tied up ZA operand for accumulate instructions
May 20 2022, 2:48 AM · Restricted Project, Restricted Project
CarolineConcatto closed D125537: [AArch64][SME]Tied up ZA operand for accumulate instructions.
May 20 2022, 2:48 AM · Restricted Project, Restricted Project
CarolineConcatto committed rG8765ad42cd01: [AArch64][SME][NFC] Add implicit operands for SME instructions in the… (authored by CarolineConcatto).
[AArch64][SME][NFC] Add implicit operands for SME instructions in the…
May 20 2022, 2:41 AM · Restricted Project, Restricted Project
CarolineConcatto closed D125534: [AArch64][SME][NFC] Add implicit operands for SME instructions in the disassembly..
May 20 2022, 2:41 AM · Restricted Project, Restricted Project

May 13 2022

CarolineConcatto added reviewers for D125534: [AArch64][SME][NFC] Add implicit operands for SME instructions in the disassembly.: sdesmalen, c-rhodes, pratlucas.
May 13 2022, 4:22 AM · Restricted Project, Restricted Project
CarolineConcatto added reviewers for D125537: [AArch64][SME]Tied up ZA operand for accumulate instructions: sdesmalen, c-rhodes, david-arm.
May 13 2022, 4:19 AM · Restricted Project, Restricted Project
CarolineConcatto requested review of D125537: [AArch64][SME]Tied up ZA operand for accumulate instructions.
May 13 2022, 4:14 AM · Restricted Project, Restricted Project
CarolineConcatto requested review of D125534: [AArch64][SME][NFC] Add implicit operands for SME instructions in the disassembly..
May 13 2022, 3:45 AM · Restricted Project, Restricted Project

Mar 11 2022

CarolineConcatto updated the diff for D120894: [AArch64][SVE]Make better use of gather/scatter when inside a loop body.

Removed unnecessary for loop from isInstUsedByGEP

Mar 11 2022, 5:59 AM · Restricted Project, Restricted Project

Mar 10 2022

CarolineConcatto added inline comments to D120894: [AArch64][SVE]Make better use of gather/scatter when inside a loop body.
Mar 10 2022, 7:22 AM · Restricted Project, Restricted Project
CarolineConcatto updated the diff for D120894: [AArch64][SVE]Make better use of gather/scatter when inside a loop body.
  • Earlier return in isInstUsedByGEP when I has more than 1 use
  • Update for loop in shouldSinkOperands to check all operands
Mar 10 2022, 7:20 AM · Restricted Project, Restricted Project

Mar 7 2022

CarolineConcatto added inline comments to D120894: [AArch64][SVE]Make better use of gather/scatter when inside a loop body.
Mar 7 2022, 12:28 AM · Restricted Project, Restricted Project
CarolineConcatto updated the diff for D120894: [AArch64][SVE]Make better use of gather/scatter when inside a loop body.
  • use a threshold for isInstUsedByGEP to reduce future costs.
  • add test for the threshold feature.
  • pull step and splat only when being used by a ge.p
Mar 7 2022, 12:28 AM · Restricted Project, Restricted Project

Mar 3 2022

CarolineConcatto added reviewers for D120894: [AArch64][SVE]Make better use of gather/scatter when inside a loop body: kmclaughlin, sdesmalen, david-arm.
Mar 3 2022, 5:19 AM · Restricted Project, Restricted Project
CarolineConcatto requested review of D120894: [AArch64][SVE]Make better use of gather/scatter when inside a loop body.
Mar 3 2022, 5:18 AM · Restricted Project, Restricted Project

Feb 17 2022

CarolineConcatto added inline comments to D119728: [AArch64][SVE] Handle more cases in findMoreOptimalIndexType..
Feb 17 2022, 7:29 AM · Restricted Project

Feb 3 2022

CarolineConcatto committed rG961e954af592: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices (authored by CarolineConcatto).
[AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices
Feb 3 2022, 11:23 AM
CarolineConcatto closed D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
Feb 3 2022, 11:23 AM · Restricted Project
CarolineConcatto committed rG019f0221d52d: [AArch64][SVE] Fold gather/scatter with 32bits when possible (authored by CarolineConcatto).
[AArch64][SVE] Fold gather/scatter with 32bits when possible
Feb 3 2022, 11:00 AM
CarolineConcatto closed D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Feb 3 2022, 11:00 AM · Restricted Project
CarolineConcatto updated the diff for D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
  • rebase on top of main
Feb 3 2022, 10:48 AM · Restricted Project
CarolineConcatto updated the diff for D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.

-Remove commented line from test gather_i8_index_offset_8

Feb 3 2022, 9:58 AM · Restricted Project
CarolineConcatto updated the diff for D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
  • Update test gather_i8_index_offset_8
Feb 3 2022, 9:55 AM · Restricted Project
CarolineConcatto added inline comments to D118917: [AArch64][NFC] Remove call to useSVEForFixedLengthVectors in useSVEForFixedLengthVectorVT.
Feb 3 2022, 9:01 AM · Restricted Project
CarolineConcatto updated the diff for D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
  • Create SDValues Add, ShiftOp ,StepOp ,OffsetOp as requested
Feb 3 2022, 8:06 AM · Restricted Project

Feb 2 2022

CarolineConcatto updated the diff for D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
  • Use dyn_cast_or_null<ConstantSDNode> to create Shift
  • get Shift value with SDValue(Shift, 0)

as suggested by Sander

Feb 2 2022, 8:19 AM · Restricted Project
CarolineConcatto added inline comments to D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
Feb 2 2022, 5:50 AM · Restricted Project
CarolineConcatto updated the diff for D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
  • Address nits
Feb 2 2022, 5:50 AM · Restricted Project
CarolineConcatto added a comment to D118757: [AArch64] Remove unused feature flags from AArch64TargetInfo.

Just passing by, but I was wondering when I saw this patch.
Why you did not remove this also from lib/Driver/ToolChains/Arch/AArch64.cpp?
And the ones in llvm (AArch64.td, AArch64InstrInfo.td)?

Feb 2 2022, 2:34 AM · Restricted Project

Feb 1 2022

CarolineConcatto accepted D118696: Fix incorrect TypeSize->uint64_t cast in InductionDescriptor::isInductionPHI.

It looks like a sensible patch.
I don't see a reason why it should not be approved as it is.

Feb 1 2022, 5:40 AM · Restricted Project
CarolineConcatto added a comment to D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.

The commit message was updated, problems is that I forgot to do arc diff --verbatim.
So it does not update the GUI here. I will use next time.

Feb 1 2022, 2:37 AM · Restricted Project
CarolineConcatto updated the diff for D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
  • Add gather test
Feb 1 2022, 2:36 AM · Restricted Project

Jan 31 2022

CarolineConcatto added a comment to D118561: [AArch64] Add legal types for Streaming SVE.

It looks good to me. But I would let someone else give the final decision.
I don't know what are the other consequences of having this merged.
But it looks a reasonable patch. I believe having this approved means fixing the tests in :
https://reviews.llvm.org/D118562

Jan 31 2022, 9:39 AM · Restricted Project, Restricted Project
CarolineConcatto added a comment to D118562: [AArch64] Allow copying of SVE registers in Streaming SVE.

Just to check, this change depends on https://reviews.llvm.org/D118561. Correct?
That is the reason your tests are failing for now.
LLVM ERROR: Don't know how to legalize this scalable vector type
I think you need to use git-clang-format.

Jan 31 2022, 9:35 AM · Restricted Project, Restricted Project
CarolineConcatto updated the diff for D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
  • change commit message tittle
  • rebase depends on D118459 too
Jan 31 2022, 9:25 AM · Restricted Project
CarolineConcatto updated the summary of D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Jan 31 2022, 9:01 AM · Restricted Project
CarolineConcatto added inline comments to D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Jan 31 2022, 9:00 AM · Restricted Project
CarolineConcatto updated the diff for D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
  • Address Sander's comments
Jan 31 2022, 9:00 AM · Restricted Project

Jan 28 2022

CarolineConcatto added inline comments to D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Jan 28 2022, 6:43 AM · Restricted Project
CarolineConcatto updated the diff for D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
  • Rebase the patch under D118459 and address comments
Jan 28 2022, 6:42 AM · Restricted Project

Jan 27 2022

CarolineConcatto edited reviewers for D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices, added: sdesmalen; removed: SanderSpies.
Jan 27 2022, 3:27 AM · Restricted Project
CarolineConcatto added inline comments to D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Jan 27 2022, 3:26 AM · Restricted Project
CarolineConcatto added reviewers for D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices: SanderSpies, david-arm, kmclaughlin.
Jan 27 2022, 3:26 AM · Restricted Project
CarolineConcatto requested review of D118345: [AArch64][SVE] Add more folds to make use of gather/scatter with 32-bit indices.
Jan 27 2022, 3:25 AM · Restricted Project
CarolineConcatto added inline comments to D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Jan 27 2022, 3:24 AM · Restricted Project
CarolineConcatto updated the diff for D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
  • This patch has only this pattern GEP (%ptr, (splat(%offset) + stepvector(A)))
Jan 27 2022, 3:24 AM · Restricted Project

Jan 21 2022

CarolineConcatto updated the summary of D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Jan 21 2022, 9:09 AM · Restricted Project
CarolineConcatto requested review of D117900: [AArch64][SVE] Fold gather/scatter with 32bits when possible.
Jan 21 2022, 9:05 AM · Restricted Project
CarolineConcatto committed rGad43217a0466: [InstCombine] Fold for masked gather when loading the same value each time. (authored by CarolineConcatto).
[InstCombine] Fold for masked gather when loading the same value each time.
Jan 21 2022, 6:20 AM
CarolineConcatto closed D115726: [InstCombine] Fold for masked gather when loading the same value each time..
Jan 21 2022, 6:20 AM · Restricted Project

Jan 14 2022

CarolineConcatto updated the diff for D115726: [InstCombine] Fold for masked gather when loading the same value each time..
  • Remove test negative_gather_v2i64_uniform_ptrs_no_all_active_mask
Jan 14 2022, 2:06 AM · Restricted Project
CarolineConcatto committed rG8e5a5b619d34: [InstCombine] Fold for masked scatters to a uniform address (authored by CarolineConcatto).
[InstCombine] Fold for masked scatters to a uniform address
Jan 14 2022, 1:56 AM
CarolineConcatto closed D115724: [InstCombine] Fold for masked scatters to a uniform address.
Jan 14 2022, 1:56 AM · Restricted Project

Jan 7 2022

CarolineConcatto added inline comments to D115726: [InstCombine] Fold for masked gather when loading the same value each time..
Jan 7 2022, 2:01 AM · Restricted Project
CarolineConcatto updated the diff for D115726: [InstCombine] Fold for masked gather when loading the same value each time..
  • Improve comments in the test file
Jan 7 2022, 2:01 AM · Restricted Project
CarolineConcatto added inline comments to D115724: [InstCombine] Fold for masked scatters to a uniform address.
Jan 7 2022, 1:24 AM · Restricted Project
CarolineConcatto updated the diff for D115724: [InstCombine] Fold for masked scatters to a uniform address.
  • Improve comments in the test file
Jan 7 2022, 1:23 AM · Restricted Project

Jan 6 2022

CarolineConcatto updated the diff for D115726: [InstCombine] Fold for masked gather when loading the same value each time..
  • remove sve attribute from masked_intrinsics.ll
  • remove negative test with vscale
Jan 6 2022, 1:36 AM · Restricted Project
CarolineConcatto updated the diff for D115724: [InstCombine] Fold for masked scatters to a uniform address.
  • Remove sve attribute from masked_intrinsics.ll
Jan 6 2022, 1:11 AM · Restricted Project

Jan 4 2022

CarolineConcatto updated the diff for D115724: [InstCombine] Fold for masked scatters to a uniform address.
  • Move positive tests in vscale_masked_intrinsics.ll to masked_intrinsics.ll
  • Rename tests as suggested in D115726
  • Fix test scatter_nxv4i16_uniform_vals_uniform_ptrs_all_active_mask
Jan 4 2022, 7:39 AM · Restricted Project
CarolineConcatto added a comment to D115726: [InstCombine] Fold for masked gather when loading the same value each time..

Thank you Sander!

Jan 4 2022, 6:04 AM · Restricted Project
CarolineConcatto updated the diff for D115726: [InstCombine] Fold for masked gather when loading the same value each time..
  • Move tests from vscale_masked_intrinsics.ll to masked_intrinsics.ll .
  • Add test when mask is all zero
Jan 4 2022, 6:04 AM · Restricted Project

Dec 17 2021

CarolineConcatto added a comment to D115726: [InstCombine] Fold for masked gather when loading the same value each time..

Thank you @david-arm for the suggestions.
I believe I have addressed all of them.

Dec 17 2021, 6:19 AM · Restricted Project
CarolineConcatto updated the diff for D115726: [InstCombine] Fold for masked gather when loading the same value each time..

-address comments in the test

Dec 17 2021, 6:18 AM · Restricted Project
CarolineConcatto added inline comments to D115724: [InstCombine] Fold for masked scatters to a uniform address.
Dec 17 2021, 5:57 AM · Restricted Project
CarolineConcatto updated the diff for D115724: [InstCombine] Fold for masked scatters to a uniform address.

-Changed the mask to be an input for vscale test when it is
not all active or inactive

Dec 17 2021, 5:57 AM · Restricted Project

Dec 16 2021

CarolineConcatto added inline comments to D115724: [InstCombine] Fold for masked scatters to a uniform address.
Dec 16 2021, 7:25 AM · Restricted Project
CarolineConcatto updated the diff for D115724: [InstCombine] Fold for masked scatters to a uniform address.
  • Address comments in the test
  • Remove redundant comment in the function
Dec 16 2021, 7:25 AM · Restricted Project
CarolineConcatto added a comment to D115724: [InstCombine] Fold for masked scatters to a uniform address.

Thank you @sdesmalen and @david-arm.
I believe both folds:

  1. The pointer is a splat(), the value is not a splat, the mask is an all-active mask.
  2. The pointer is a splat(), the value is a splat(), the mask is not all-inactive (but not necessarily all active). are now implemented.

I noticed that for option 2 (when the value is a splat) in the extract element the value is the same as the original value that was splatted.
So in case 2, the fold is to a scalar store of value in the destination pointer. I don't need to extract the element from any line.
So I only have:
StoreInst *S = new StoreInst(SplatValue, SplatPtr, false, Alignment);
I also added the same tests for fixed and scalable vector.

Dec 16 2021, 5:32 AM · Restricted Project
CarolineConcatto updated the diff for D115724: [InstCombine] Fold for masked scatters to a uniform address.
  • Add fold for the value when it is a splat.
  • Add tests for fixed vector
Dec 16 2021, 5:23 AM · Restricted Project

Dec 14 2021

CarolineConcatto added reviewers for D115726: [InstCombine] Fold for masked gather when loading the same value each time.: sebastian-ne, Flakebi, kmclaughlin, david-arm, sdesmalen.
Dec 14 2021, 6:13 AM · Restricted Project
CarolineConcatto requested review of D115726: [InstCombine] Fold for masked gather when loading the same value each time..
Dec 14 2021, 6:12 AM · Restricted Project
CarolineConcatto added reviewers for D115724: [InstCombine] Fold for masked scatters to a uniform address: ctetreau, sdesmalen, david-arm, kmclaughlin.
Dec 14 2021, 5:52 AM · Restricted Project
CarolineConcatto requested review of D115724: [InstCombine] Fold for masked scatters to a uniform address.
Dec 14 2021, 5:49 AM · Restricted Project

Oct 28 2021

CarolineConcatto committed rG2186b011e966: [Driver][AArch64]Add driver support for neoverse-512tvb target (authored by CarolineConcatto).
[Driver][AArch64]Add driver support for neoverse-512tvb target
Oct 28 2021, 1:10 AM
CarolineConcatto closed D112406: [Driver][AArch64]Add driver support for neoverse-512tvb target.
Oct 28 2021, 1:10 AM · Restricted Project, Restricted Project

Oct 27 2021

CarolineConcatto committed rG1137b7207d70: [SelectionDAG] Widening the result of INSERT_SUBVECTOR. (authored by CarolineConcatto).
[SelectionDAG] Widening the result of INSERT_SUBVECTOR.
Oct 27 2021, 5:54 AM
CarolineConcatto closed D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR.
Oct 27 2021, 5:54 AM · Restricted Project
CarolineConcatto retitled D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR from [SelectionDAG] Add widen result function for insert subvector to [SelectionDAG] Widening the result of INSERT_SUBVECTOR.
Oct 27 2021, 5:10 AM · Restricted Project
CarolineConcatto added inline comments to D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR.
Oct 27 2021, 2:32 AM · Restricted Project
CarolineConcatto updated the diff for D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR.
  • Fix insert_nxv6i32_nxv2i32 to use sv0 and sv1 as input instead of only sv0
Oct 27 2021, 2:31 AM · Restricted Project

Oct 26 2021

CarolineConcatto added a comment to D112406: [Driver][AArch64]Add driver support for neoverse-512tvb target.

Thank you for your review @dmgreen and @tschuett.
I rebase the patch, now VScaleForTuning is being set.
And I removed support for neoverse-512tvb from Arm 32 bits.
@paulwalker-arm pointed me that neoverse-512tvb is only supported by AArch64.

Oct 26 2021, 5:25 AM · Restricted Project, Restricted Project
CarolineConcatto updated the diff for D112406: [Driver][AArch64]Add driver support for neoverse-512tvb target.
  • Rebase and remove support on for ARM 32 bits
Oct 26 2021, 5:19 AM · Restricted Project, Restricted Project

Oct 25 2021

CarolineConcatto accepted D112320: [SVE][CodeGen] Fix incorrect legalisation of zero-extended masked loads.

Thank you @kmclaughlin

Oct 25 2021, 8:21 AM · Restricted Project
CarolineConcatto retitled D112406: [Driver][AArch64]Add driver support for neoverse-512tvb target from [WIP][Driver][AArch64]Add driver support for neoverse-512tvb target to [Driver][AArch64]Add driver support for neoverse-512tvb target.
Oct 25 2021, 8:03 AM · Restricted Project, Restricted Project
CarolineConcatto updated the diff for D112406: [Driver][AArch64]Add driver support for neoverse-512tvb target.
  • remove neoverse-512tvb from Host.cpp
Oct 25 2021, 8:01 AM · Restricted Project, Restricted Project
CarolineConcatto retitled D112406: [Driver][AArch64]Add driver support for neoverse-512tvb target from [Driver][AArch64]Add driver support for neoverse-512tvb target to [WIP][Driver][AArch64]Add driver support for neoverse-512tvb target.
Oct 25 2021, 5:09 AM · Restricted Project, Restricted Project

Oct 24 2021

CarolineConcatto requested review of D112406: [Driver][AArch64]Add driver support for neoverse-512tvb target.
Oct 24 2021, 11:59 PM · Restricted Project, Restricted Project

Oct 22 2021

CarolineConcatto added a comment to D112320: [SVE][CodeGen] Fix incorrect legalisation of zero-extended masked loads.

Hey @kmclaughlin,
Thank you for adding me as a reviewer of the patch.
Do we know what was the logic to set ExtType to ISD::EXTLOAD?
I don't see that on other parts of the legalisation for Masked load.

Oct 22 2021, 9:00 AM · Restricted Project

Oct 21 2021

CarolineConcatto updated the diff for D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR.
  • Add new tests insert_nxv3i32_nxv2i32_2 and insert_nxv3f32_nxv2f32
Oct 21 2021, 2:28 AM · Restricted Project
CarolineConcatto updated the diff for D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR.
  • Fix return value in the test insert_nxv6i32_nxv2i32
Oct 21 2021, 1:37 AM · Restricted Project
CarolineConcatto edited reviewers for D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR, added: dmgreen; removed: greened.
Oct 21 2021, 1:24 AM · Restricted Project

Oct 20 2021

CarolineConcatto added reviewers for D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR: sdesmalen, greened, kmclaughlin, david-arm.
Oct 20 2021, 3:09 PM · Restricted Project
CarolineConcatto requested review of D112187: [SelectionDAG] Widening the result of INSERT_SUBVECTOR.
Oct 20 2021, 3:07 PM · Restricted Project
CarolineConcatto added a comment to D112078: [AArch64][SVE] Add new ld<n> intrinsics that return a struct of vscale types.

Hi @bsmith
I was not added to the review so you can ignore my comments.
But if you want to increase a fellow contributor knowledge in LLVM I posted some questions.

Oct 20 2021, 12:20 AM · Restricted Project

Oct 12 2021

CarolineConcatto added inline comments to D111633: [SelectionDAG] Fix getVectorSubVecPointer for scalable subvectors..
Oct 12 2021, 5:18 AM · Restricted Project

Sep 29 2021

CarolineConcatto added inline comments to D110624: [SVE] Fix incorrect DAG combines when extracting fixed-width from scalable vectors.
Sep 29 2021, 7:08 AM · Restricted Project

Sep 20 2021

CarolineConcatto accepted D109808: [InstCombine] Eliminate vector reverse if all inputs/outputs to an instruction are reverses.

Thank you @mnadeem.
For me, this patch looks good to land on main.

Sep 20 2021, 5:38 AM · Restricted Project

Sep 16 2021

CarolineConcatto added a comment to D109808: [InstCombine] Eliminate vector reverse if all inputs/outputs to an instruction are reverses.

@mnadeem Nice patch.
I believe you need to test the other way around for binary operations, no?
I left a comment, I hope it makes sense.

Sep 16 2021, 8:55 AM · Restricted Project

Sep 9 2021

CarolineConcatto accepted D109313: [SelectionDAG] PromoteIntRes_EXTRACT_SUBVECTOR for scalable vectors..

Thank you Sander!

Sep 9 2021, 6:15 AM · Restricted Project