Page MenuHomePhabricator

CarolineConcatto (Caroline)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 4 2020, 3:58 AM (62 w, 1 d)

Recent Activity

Today

CarolineConcatto added reviewers for D100485: [AArch64][SVE] Fix crash with icmp+select: sdesmalen, c-rhodes, bsmith, joechrisellis.
Wed, Apr 14, 8:09 AM · Restricted Project
CarolineConcatto requested review of D100485: [AArch64][SVE] Fix crash with icmp+select.
Wed, Apr 14, 8:06 AM · Restricted Project

Thu, Apr 1

CarolineConcatto accepted D99192: [NFC] Add tests for scalable vectorization of loops with large stride acesses.

LGMT David!
Learning new things every day!

Thu, Apr 1, 2:24 AM · Restricted Project
CarolineConcatto added inline comments to D99192: [NFC] Add tests for scalable vectorization of loops with large stride acesses.
Thu, Apr 1, 1:24 AM · Restricted Project

Wed, Mar 31

CarolineConcatto added a comment to D99192: [NFC] Add tests for scalable vectorization of loops with large stride acesses.

Hey David,
I believe I have some nit. :)

Wed, Mar 31, 11:55 AM · Restricted Project

Fri, Mar 26

CarolineConcatto added inline comments to D98934: [SVE][CostModel] Add instruction cost for operations on scalable vectors.
Fri, Mar 26, 6:11 AM · Restricted Project

Thu, Mar 18

CarolineConcatto added a comment to D98351: [InstCombine] shrinkFPConstantVector(): scalable vector support.

Just in case ...

Thu, Mar 18, 10:47 AM · Restricted Project

Tue, Mar 16

CarolineConcatto committed rG3c03635d5300: [SVE][LoopVectorize] Add support for scalable vectorization of loops with… (authored by CarolineConcatto).
[SVE][LoopVectorize] Add support for scalable vectorization of loops with…
Tue, Mar 16, 2:19 AM
CarolineConcatto closed D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Tue, Mar 16, 2:19 AM · Restricted Project

Mar 15 2021

CarolineConcatto added a comment to D98506: [SVE][LoopVectorize] Verify support for vectorizing loops with invariant loads.

Hey Kerry,

Mar 15 2021, 2:19 AM · Restricted Project
CarolineConcatto added a comment to D98487: [AArch64][SVE/NEON] Add support for FROUNDEVEN for both NEON and fixed length SVE.

Thank you for adding me as a reviewer, although I don't think I am the more qualified to approve or not this patch.
But I have a question:
Why is this patch only changing int_aarch64_neon_frintn and not int_aarch64_sve_frintn?
Is there a particular reason to do so?
As you said in the commit message the ISD node for FROUNDEVEN exists now.
If so it would be too much to explain that in the commit message?
Thank you,
Carol

Mar 15 2021, 2:03 AM · Restricted Project, Restricted Project

Mar 10 2021

CarolineConcatto added a comment to D98351: [InstCombine] shrinkFPConstantVector(): scalable vector support.

Hey Nash,

Mar 10 2021, 11:50 PM · Restricted Project

Mar 8 2021

CarolineConcatto added a comment to D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.

Hey @david-arm sorry for the nit in the tests.
Usually happens when I do copy and paste many times.
I've updated the patch and rebased it with the main.

Mar 8 2021, 5:37 AM · Restricted Project
CarolineConcatto updated the diff for D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
  • Fix nit in the .ll tests and LoopVectorize.cpp
  • Remote copied getRunTimeVF from LoopVectorize.cpp
  • Move CreateVectorReverse implementation from IRBuilder.h to IRBuilder.cpp
Mar 8 2021, 5:34 AM · Restricted Project

Mar 5 2021

CarolineConcatto added inline comments to D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Mar 5 2021, 1:42 AM · Restricted Project
CarolineConcatto added a comment to D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.

Thank you all for the review.
I C&P the function getRunTime from D95139.
If D95139 is merged before this patch I will remove it.
All the other dependencies are already in main, and getRunTimeVF is the only function that this patch needs from D95139.

Mar 5 2021, 1:36 AM · Restricted Project
CarolineConcatto updated the diff for D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
  • Add tests for when vector.reverse needs to reverse a mask
  • improve code in LoopVectorize.cpp
  • C&P function getRunTimeVF from D95139
Mar 5 2021, 1:22 AM · Restricted Project

Mar 4 2021

CarolineConcatto committed rGf2b749be1568: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector (authored by CarolineConcatto).
[CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector
Mar 4 2021, 10:54 AM
CarolineConcatto closed D97758: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector.
Mar 4 2021, 10:54 AM · Restricted Project

Mar 3 2021

CarolineConcatto edited reviewers for D97758: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector, added: fhahn; removed: Florian.
Mar 3 2021, 12:33 AM · Restricted Project

Mar 2 2021

CarolineConcatto added a comment to D97758: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector.

Thank you Sander!
My mistake not update main. Thank you for pointing it out.
Now the test is based on the latest main repo.

Mar 2 2021, 8:47 AM · Restricted Project
CarolineConcatto updated the diff for D97758: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector.
  • return file getIntrinsicInstrCost-vector-reverse.ll with the i1 test
Mar 2 2021, 8:44 AM · Restricted Project
CarolineConcatto updated the diff for D97758: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector.
  • update the llvm-ir test
  • rebase
Mar 2 2021, 8:18 AM · Restricted Project
CarolineConcatto retitled D97758: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector from [AArch64] Add cost model for shuffle reverse with i1 and scalable vector to [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector.
Mar 2 2021, 1:38 AM · Restricted Project
CarolineConcatto requested review of D97758: [CostModel][SVE] Add cost model for shuffle reverse with i1 and scalable vector.
Mar 2 2021, 1:36 AM · Restricted Project

Feb 15 2021

CarolineConcatto committed rG99dbc0fa7601: [LangRef] Increase size of title underline for experimental.vector.reverse (authored by CarolineConcatto).
[LangRef] Increase size of title underline for experimental.vector.reverse
Feb 15 2021, 7:26 AM
CarolineConcatto committed rGb52e6c58911f: [CostModel]Add cost model for experimental.vector.reverse (authored by CarolineConcatto).
[CostModel]Add cost model for experimental.vector.reverse
Feb 15 2021, 6:26 AM
CarolineConcatto closed D95603: [CostModel]Add cost model for experimental.vector.reverse.
Feb 15 2021, 6:25 AM · Restricted Project
CarolineConcatto committed rG2d728bbff5c6: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse (authored by CarolineConcatto).
[CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse
Feb 15 2021, 5:41 AM
CarolineConcatto closed D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.
Feb 15 2021, 5:41 AM · Restricted Project

Feb 11 2021

CarolineConcatto added a comment to D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.

Thank you @david-arm.
I've removed the pre.header.
It is indeed better.

Feb 11 2021, 5:38 AM · Restricted Project
CarolineConcatto updated the diff for D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
  • change index 'i' in the tests to be 64 bits instead of 32bits
Feb 11 2021, 5:36 AM · Restricted Project
CarolineConcatto updated the diff for D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.

-removing redundant lines in the test

Feb 11 2021, 4:55 AM · Restricted Project
CarolineConcatto updated the diff for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.

-change the names of the tests to named-vector-shuffle-reverse
-add text in LangRef about the use of experimental.vector.reverse

Feb 11 2021, 4:42 AM · Restricted Project

Feb 10 2021

CarolineConcatto added inline comments to D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Feb 10 2021, 12:16 PM · Restricted Project
CarolineConcatto updated the diff for D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
  • change how to get NumElt inside LoopVectorize
  • change the index to be 64 bits
Feb 10 2021, 12:05 PM · Restricted Project

Feb 9 2021

CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Feb 9 2021, 9:08 AM · Restricted Project
CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Feb 9 2021, 9:07 AM · Restricted Project
CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Feb 9 2021, 9:05 AM · Restricted Project

Feb 7 2021

CarolineConcatto added a comment to D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.

Thank you all for the review.
I believe I had addressed your comments.
I've created another patch to compute the cost for vector reverse with scalable vector.
I've removed the cast from int to double in one of the tests, so we don't need cast fixed for this feature.
And the IR builder now has a reverse function for fixed and scalable vectors.

Feb 7 2021, 10:49 AM · Restricted Project
CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Feb 7 2021, 10:43 AM · Restricted Project
CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Feb 7 2021, 10:42 AM · Restricted Project
CarolineConcatto updated the diff for D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
  • Move cost function for Shuffle reverse to D93639
  • Use getRunTimeVF to compute the runtime vector size for scalable vector
  • add tests for fixed vector
Feb 7 2021, 10:41 AM · Restricted Project

Feb 3 2021

CarolineConcatto committed rG2cbcf3e297d1: [AArch64][SVE]Add cost model for broadcast shuffle (authored by CarolineConcatto).
[AArch64][SVE]Add cost model for broadcast shuffle
Feb 3 2021, 1:56 AM
CarolineConcatto closed D95598: [AArch64][SVE]Add cost model for broadcast shuffle.
Feb 3 2021, 1:56 AM · Restricted Project
CarolineConcatto updated the diff for D95598: [AArch64][SVE]Add cost model for broadcast shuffle.

-remove non used parameter from the test function broadcast

Feb 3 2021, 1:21 AM · Restricted Project

Jan 29 2021

CarolineConcatto added a comment to D95603: [CostModel]Add cost model for experimental.vector.reverse.

Thank you @david-arm for the test suggestion.
I also added the support for nxv8bf16.

Jan 29 2021, 9:16 AM · Restricted Project
CarolineConcatto updated the diff for D95603: [CostModel]Add cost model for experimental.vector.reverse.
  • add support for nxv8bf16 and simplify .ll test
Jan 29 2021, 9:14 AM · Restricted Project
CarolineConcatto added a comment to D95598: [AArch64][SVE]Add cost model for broadcast shuffle.

@Thank you all for the suggestion.
I have added the support for MVT::nxv8bf16 and simplify the test.

Jan 29 2021, 8:24 AM · Restricted Project
CarolineConcatto updated the diff for D95598: [AArch64][SVE]Add cost model for broadcast shuffle.
  • correct style
Jan 29 2021, 8:22 AM · Restricted Project
CarolineConcatto updated the diff for D95598: [AArch64][SVE]Add cost model for broadcast shuffle.
  • add bfloat test and support
Jan 29 2021, 8:19 AM · Restricted Project
CarolineConcatto updated the diff for D95598: [AArch64][SVE]Add cost model for broadcast shuffle.
  • simplify sve broadcast test
Jan 29 2021, 7:46 AM · Restricted Project
CarolineConcatto added inline comments to D94230: [AArch64][SVE] Coalesce ptrue instrinsic calls where possible.
Jan 29 2021, 1:58 AM · Restricted Project

Jan 28 2021

CarolineConcatto added inline comments to D95598: [AArch64][SVE]Add cost model for broadcast shuffle.
Jan 28 2021, 3:10 AM · Restricted Project
CarolineConcatto updated the diff for D95598: [AArch64][SVE]Add cost model for broadcast shuffle.

-replace test name @slapt_v4f32 by @broadcast_v4f32(

Jan 28 2021, 3:04 AM · Restricted Project
CarolineConcatto updated the diff for D95598: [AArch64][SVE]Add cost model for broadcast shuffle.

-add original if statement
-move the table cost for scalable types to ShuffleTbl

Jan 28 2021, 3:01 AM · Restricted Project
CarolineConcatto added inline comments to D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.
Jan 28 2021, 2:49 AM · Restricted Project
CarolineConcatto updated the diff for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.

-remove test for f16 on X86 as they are not legal.

Jan 28 2021, 2:47 AM · Restricted Project
CarolineConcatto added a reviewer for D95598: [AArch64][SVE]Add cost model for broadcast shuffle: david-arm.
Jan 28 2021, 2:42 AM · Restricted Project
CarolineConcatto added reviewers for D95603: [CostModel]Add cost model for experimental.vector.reverse: sdesmalen, david-arm, paulwalker-arm, fhahn, cameron.mcinally, frasercrmck.
Jan 28 2021, 2:31 AM · Restricted Project
CarolineConcatto added reviewers for D95598: [AArch64][SVE]Add cost model for broadcast shuffle: sdesmalen, samparker, RKSimon.
Jan 28 2021, 2:29 AM · Restricted Project
CarolineConcatto requested review of D95603: [CostModel]Add cost model for experimental.vector.reverse.
Jan 28 2021, 2:02 AM · Restricted Project
CarolineConcatto requested review of D95598: [AArch64][SVE]Add cost model for broadcast shuffle.
Jan 28 2021, 1:05 AM · Restricted Project

Jan 27 2021

CarolineConcatto added a comment to D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.

Thank you all for the comments.
I've added tests for -O0 in AArch64 and fixed-width test for X86.
ATM the X86 test has no specific target feature.

Jan 27 2021, 12:43 AM · Restricted Project

Jan 26 2021

CarolineConcatto updated the diff for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.

-add -O0 and X86 test for vector.reverse

Jan 26 2021, 9:03 AM · Restricted Project

Jan 25 2021

CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Jan 25 2021, 7:52 AM · Restricted Project
CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Jan 25 2021, 7:50 AM · Restricted Project
CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Jan 25 2021, 7:49 AM · Restricted Project
CarolineConcatto updated the summary of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Jan 25 2021, 7:31 AM · Restricted Project
CarolineConcatto updated the diff for D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.

-s/scalabe/scalable

Jan 25 2021, 7:31 AM · Restricted Project
CarolineConcatto retitled D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse from [SVELoopVectorize] Add support for scalable vectorization of loops with vector reverse to [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Jan 25 2021, 7:27 AM · Restricted Project
CarolineConcatto requested review of D95363: [SVE][LoopVectorize] Add support for scalable vectorization of loops with vector reverse.
Jan 25 2021, 7:22 AM · Restricted Project
CarolineConcatto added a comment to D95245: [SVE] Add support for scalable vectorization of loops with int/fast FP reductions.

Hey Kerry,
Thank you for this patch.
I found some nit and I have some suggestions about instructionCost.

Jan 25 2021, 1:18 AM · Restricted Project

Jan 21 2021

CarolineConcatto added a reviewer for D94708: [IR] Introduce llvm.experimental.vector.splice intrinsic: CarolineConcatto.
Jan 21 2021, 8:51 AM · Restricted Project

Jan 19 2021

CarolineConcatto added inline comments to D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.
Jan 19 2021, 8:39 AM · Restricted Project
CarolineConcatto updated the diff for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.

-fix table gen style for TargetSelectionDAG.td

Jan 19 2021, 8:37 AM · Restricted Project
CarolineConcatto updated the diff for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.

-fix style on LegalizeIntegerTypes.cpp

Jan 19 2021, 8:24 AM · Restricted Project
CarolineConcatto updated the diff for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.

-remove AArch64 custom lowering
-move SNode vector_reverse to TargetSelectionDAG.td

Jan 19 2021, 8:21 AM · Restricted Project
CarolineConcatto updated the summary of D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.
Jan 19 2021, 7:56 AM · Restricted Project
CarolineConcatto committed rG172f1f8952c9: [AArch64][SVE]Add cost model for vector reduce for scalable vector (authored by CarolineConcatto).
[AArch64][SVE]Add cost model for vector reduce for scalable vector
Jan 19 2021, 3:55 AM
CarolineConcatto closed D93639: [AArch64][SVE]Add cost model for vector reduce for scalable vector.
Jan 19 2021, 3:54 AM · Restricted Project

Jan 18 2021

CarolineConcatto accepted D92238: [SCEVExpander] Migrate costAndCollectOperands to use InstructionCost..

It look to me now.

Jan 18 2021, 6:42 AM · Restricted Project
CarolineConcatto added reviewers for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse: rogfer01, cameron.mcinally.
Jan 18 2021, 6:28 AM · Restricted Project
CarolineConcatto committed rG36710c38c1b7: [NFC]Migrate VectorCombine.cpp to use InstructionCost (authored by CarolineConcatto).
[NFC]Migrate VectorCombine.cpp to use InstructionCost
Jan 18 2021, 5:38 AM
CarolineConcatto closed D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.
Jan 18 2021, 5:37 AM · Restricted Project
CarolineConcatto updated the diff for D93639: [AArch64][SVE]Add cost model for vector reduce for scalable vector.

-re-arrange the test
-remove f<operand> tests with integer

Jan 18 2021, 2:33 AM · Restricted Project

Jan 17 2021

CarolineConcatto updated the diff for D93639: [AArch64][SVE]Add cost model for vector reduce for scalable vector.
  • change vector size to 128bits
Jan 17 2021, 2:12 PM · Restricted Project
CarolineConcatto added reviewers for D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse: sdesmalen, david-arm, paulwalker-arm, kmclaughlin.
Jan 17 2021, 11:56 AM · Restricted Project
CarolineConcatto added a comment to D93639: [AArch64][SVE]Add cost model for vector reduce for scalable vector.

@Thank you Sander,
I added more tests. Now we have tests for legal and not legal, besides the test with integer for FP operands.

Jan 17 2021, 11:55 AM · Restricted Project
CarolineConcatto updated the diff for D93639: [AArch64][SVE]Add cost model for vector reduce for scalable vector.

-add missing test for fmax,fadd with integer

Jan 17 2021, 11:55 AM · Restricted Project
CarolineConcatto requested review of D94883: [CodeGen][SelectionDAG]Add new intrinsic experimental.vector.reverse.
Jan 17 2021, 11:45 AM · Restricted Project
CarolineConcatto updated the summary of D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.
Jan 17 2021, 11:09 AM · Restricted Project
CarolineConcatto updated the diff for D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.

-replace (!NewCost.isValid && !OldCost.isvalid()) by !NewCost.isValid()

Jan 17 2021, 11:08 AM · Restricted Project
CarolineConcatto updated the diff for D93639: [AArch64][SVE]Add cost model for vector reduce for scalable vector.
  • remove cost for vector.reduce.{mul"fmul}
  • add regression tests for not legal vectors
Jan 17 2021, 10:40 AM · Restricted Project

Jan 13 2021

CarolineConcatto updated the diff for D93639: [AArch64][SVE]Add cost model for vector reduce for scalable vector.

-remove redundant code
-change getArithmeticReductionCostScalableVectorType to getArithmeticReductionCostSVE
-replace unsigned by int

Jan 13 2021, 10:18 AM · Restricted Project
CarolineConcatto added inline comments to D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.
Jan 13 2021, 9:28 AM · Restricted Project
CarolineConcatto updated the diff for D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.

-remove missed assert

Jan 13 2021, 9:22 AM · Restricted Project
CarolineConcatto updated the diff for D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.

-add missing invalid costs test for scalarizeBinopOrCmp

Jan 13 2021, 1:35 AM · Restricted Project
CarolineConcatto added a comment to D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.

Hi @ctetreau,
Thank you for the review.
So I removed all the asserts and added the earlier return if both costs are invalid, because in this case it means that the transformation is not cheap.
If only OldCost is invalid I believe we should do nothing, for the same reason I removed the test for !NewCost.isValid().

Jan 13 2021, 1:30 AM · Restricted Project
CarolineConcatto updated the diff for D94069: [NFC][InstructionCost]Migrate VectorCombine.cpp to use InstructionCost.

-remove asserts for OldCost invalid
-add return if both costs are invalid

Jan 13 2021, 1:26 AM · Restricted Project