Page MenuHomePhabricator

ftynse (Alex Zinenko)
User

Projects

User does not belong to any projects.

User Details

User Since
Jun 28 2018, 2:24 AM (82 w, 5 d)

Recent Activity

Today

ftynse committed rG8ed47b74300e: [mlir] NFC: use ValueRange in AffineToStandard conversion (authored by ftynse).
[mlir] NFC: use ValueRange in AffineToStandard conversion
Tue, Jan 28, 3:00 AM
ftynse committed rG6895a1c37e44: [mlir] NFC: use doxygen-style comments in AffineToStandard.cpp (authored by ftynse).
[mlir] NFC: use doxygen-style comments in AffineToStandard.cpp
Tue, Jan 28, 1:29 AM

Yesterday

ftynse added inline comments to D73488: [mlir] add lowering from affine.min to std.
Mon, Jan 27, 1:57 PM · Restricted Project
ftynse committed rG51ba5b528a98: [mlir] add lowering from affine.min to std (authored by ftynse).
[mlir] add lowering from affine.min to std
Mon, Jan 27, 1:39 PM
ftynse committed rG74df89f67f17: [NFC][mlir][linalg] Merge Utils/Intrinsics.h into EDSC/Intrinsics.h (authored by kernhanda).
[NFC][mlir][linalg] Merge Utils/Intrinsics.h into EDSC/Intrinsics.h
Mon, Jan 27, 1:39 PM
ftynse closed D73377: [NFC] [mlir] [linalg] Merge Utils/Intrinsics.h into EDSC/Intrinsics.h.
Mon, Jan 27, 1:39 PM · Restricted Project
ftynse closed D73488: [mlir] add lowering from affine.min to std.
Mon, Jan 27, 1:39 PM · Restricted Project
ftynse added inline comments to D73465: Add gpu::LaunchOp::addKernelArgument..
Mon, Jan 27, 1:38 PM · Restricted Project
ftynse added a comment to D73377: [NFC] [mlir] [linalg] Merge Utils/Intrinsics.h into EDSC/Intrinsics.h.

Could you plz rebase on master? It does not seem to apply anymore. And fix clang-format while you are there.

Mon, Jan 27, 8:21 AM · Restricted Project
ftynse created D73488: [mlir] add lowering from affine.min to std.
Mon, Jan 27, 8:21 AM · Restricted Project
ftynse accepted D73470: [mlir] fixed invalid LLVM intrinsics in LLVMOPs.td and llvmir-intrinsics.mlir..

Please note that we tend to prefix the header of the commit message with [mlir] so that people on llvm-commits@ can understand the changes are restricted to MLIR (many more people might have a say if actual LLVM IR intrinsics were broken)

Mon, Jan 27, 7:53 AM · Restricted Project
ftynse retitled D73470: [mlir] fixed invalid LLVM intrinsics in LLVMOPs.td and llvmir-intrinsics.mlir. from Fixed invalid LLVM intrinsics in LLVMOPs.td and llvmir-intrinsics.mlir. to [mlir] fixed invalid LLVM intrinsics in LLVMOPs.td and llvmir-intrinsics.mlir..
Mon, Jan 27, 7:46 AM · Restricted Project
ftynse committed rG84c3f05c8e3e: [mlir] Harden error propagation in LLVM import (authored by ftynse).
[mlir] Harden error propagation in LLVM import
Mon, Jan 27, 7:17 AM
ftynse committed rG07328944efb6: [mlir] LLVM import: handle constant data and array/vector aggregates (authored by ftynse).
[mlir] LLVM import: handle constant data and array/vector aggregates
Mon, Jan 27, 7:17 AM
ftynse closed D72834: [mlir] LLVM import: handle constant data and array/vector aggregates.
Mon, Jan 27, 7:16 AM · Restricted Project
ftynse closed D72839: [mlir] Harden error propagation in LLVM import.
Mon, Jan 27, 7:16 AM · Restricted Project
ftynse accepted D73471: Add tanh lowering from Standard dialect to NVVM and ROCDL..

Please update the commit message from "fixed tanh lowering" to "add tanh lowering". It was not broken (which would mandate a fix), it was never implemented to start with. Good to go otherwise.

Mon, Jan 27, 6:40 AM · Restricted Project
ftynse accepted D73348: Add lowering for loop.parallel to cfg..
Mon, Jan 27, 6:40 AM · Restricted Project
ftynse requested changes to D73470: [mlir] fixed invalid LLVM intrinsics in LLVMOPs.td and llvmir-intrinsics.mlir..

I suppose that, in general, there is no strict correspondance between the number of type suffixes in an overloaded intrinsic and the number of operands. So having one type may make sense for fmuladd, but may not make sense for some platform-specific vector select-like intrinsic with three arguments that would require two types. I would suggest renaming BinaryIntrinsicOp to BinarySameArgsIntrinsicOp if you do this change. Or even dropping it entirely and keeping an explicit definition for copysign and fmuladd since I don't see any other users of the Binary and TernaryIntrinsicOp classes.

Mon, Jan 27, 6:31 AM · Restricted Project
ftynse accepted D73465: Add gpu::LaunchOp::addKernelArgument..
Mon, Jan 27, 6:03 AM · Restricted Project
ftynse created D73468: [mlir] LLVM dialect: Generate conversions between EnumAttrCase and LLVM API.
Mon, Jan 27, 5:56 AM · Restricted Project
ftynse requested changes to D72962: [MLIR, OpenMP] Translation of OpenMP barrier construct to LLVM IR.

Thanks! I have a couple of comments.

Mon, Jan 27, 2:27 AM · Restricted Project
ftynse added a comment to D73348: Add lowering for loop.parallel to cfg..

Looks generally okay to me. I don't see the point of cloning before erasing, but can live with it pending a clean-up.

Mon, Jan 27, 2:15 AM · Restricted Project
ftynse added a comment to D72834: [mlir] LLVM import: handle constant data and array/vector aggregates.

Anything else @nicolasvasilache ?

Mon, Jan 27, 1:47 AM · Restricted Project
Herald added a reviewer for D73362: [mlir] EnumsGen: dissociate string form of integer enum from C++ symbol name: rriddle.

Thanks Alex for adding this! Have you considered adding an additional parameter (defaulting to C++ enumerant symbol) for the string representation instead of duplicating all the cases?

Mon, Jan 27, 1:47 AM · Restricted Project

Fri, Jan 24

ftynse added a reviewer for D73362: [mlir] EnumsGen: dissociate string form of integer enum from C++ symbol name: antiagainst.
Fri, Jan 24, 8:58 AM · Restricted Project
ftynse created D73362: [mlir] EnumsGen: dissociate string form of integer enum from C++ symbol name.
Fri, Jan 24, 8:58 AM · Restricted Project
ftynse committed rGb9013351936d: [mlir] Use all_of instead of a manual loop in IntrinsicGen. NFC (authored by ftynse).
[mlir] Use all_of instead of a manual loop in IntrinsicGen. NFC
Fri, Jan 24, 2:33 AM
ftynse requested changes to D73296: [mlir] Add MemRefTypeBuilder and refactor some MemRefType::get()..

Thanks, I like the simplification.

Fri, Jan 24, 1:15 AM · Restricted Project
ftynse added a comment to rGc8695ba9cdeb: Revert "[mlir] Add baseAttr to TypedArrayAttrBase.".

The change looks innocuous, I would have landed on approval, too.

Fri, Jan 24, 12:58 AM

Thu, Jan 23

ftynse accepted D73233: [mlir] Add option to use custom base class for dialect in LLVMIRIntrinsicGen..

Thanks!

Thu, Jan 23, 11:05 AM · Restricted Project
ftynse added inline comments to D73233: [mlir] Add option to use custom base class for dialect in LLVMIRIntrinsicGen..
Thu, Jan 23, 9:43 AM · Restricted Project
ftynse requested changes to D73145: [mlir][Linalg] Introduce folding patterns to remove certain MemRefCastOp.
Thu, Jan 23, 8:01 AM · Restricted Project
ftynse added a comment to D72802: [mlir] Introduce bare ptr calling convention for MemRefs in LLVM dialect.

I'm supportive of this change as long as Mehdi's and River's concerns are addressed.

Thu, Jan 23, 7:43 AM · Restricted Project
ftynse requested changes to D73233: [mlir] Add option to use custom base class for dialect in LLVMIRIntrinsicGen..

Looks good in general, I am suggesting to push it one step further.

Thu, Jan 23, 7:24 AM · Restricted Project
ftynse added a comment to D72888: [mlir] Add baseAttr to TypedArrayAttrBase..

I reverted this change since it broke dependent projects, and the modeling generally looks incorrect.

Thu, Jan 23, 6:03 AM · Restricted Project
ftynse committed rGc8695ba9cdeb: Revert "[mlir] Add baseAttr to TypedArrayAttrBase." (authored by ftynse).
Revert "[mlir] Add baseAttr to TypedArrayAttrBase."
Thu, Jan 23, 5:54 AM
ftynse added a reverting change for rGeec36909c18b: [mlir] Add baseAttr to TypedArrayAttrBase.: rGc8695ba9cdeb: Revert "[mlir] Add baseAttr to TypedArrayAttrBase.".
Thu, Jan 23, 5:54 AM

Wed, Jan 22

ftynse committed rG7984b47401f7: [mlir][orc] unbreak MLIR ExecutionEngine after ORC changes (authored by ftynse).
[mlir][orc] unbreak MLIR ExecutionEngine after ORC changes
Wed, Jan 22, 1:18 AM

Mon, Jan 20

ftynse committed rGf63f5a228f30: [mlir] clarify LangRef wording around control flow in regions (authored by ftynse).
[mlir] clarify LangRef wording around control flow in regions
Mon, Jan 20, 5:31 AM
ftynse accepted D72995: [MLIR] LLVM Dialect: add llvm.cmpxchg and improve llvm.atomicrmw custom parser.

Minor comments only.

Mon, Jan 20, 1:30 AM · Restricted Project
ftynse accepted D72129: [mlir] Add in-dialect lowering of gpu.all_reduce..
Mon, Jan 20, 1:21 AM · Restricted Project

Fri, Jan 17

ftynse committed rG60a0c612df51: [MLIR] LLVM dialect: Add llvm.atomicrmw (authored by flaub).
[MLIR] LLVM dialect: Add llvm.atomicrmw
Fri, Jan 17, 12:19 PM
ftynse committed rG37e2560d3d24: [Flang][mlir] add a band-aid to support the creation of mutually recursive… (authored by schweitz).
[Flang][mlir] add a band-aid to support the creation of mutually recursive…
Fri, Jan 17, 12:19 PM
ftynse closed D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.
Fri, Jan 17, 12:19 PM · Restricted Project
ftynse closed D72542: [Flang][mlir] add a band-aid to support the creation of mutually recursive types when lowering to LLVM IR.
Fri, Jan 17, 12:18 PM · Restricted Project, Restricted Project
ftynse accepted D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.
Fri, Jan 17, 12:09 PM · Restricted Project
ftynse updated subscribers of D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.

@tra, your rule adds you as blocking on all of MLIR :)

Fri, Jan 17, 12:09 PM · Restricted Project
ftynse added a comment to D72942: [mlir][Linalg] Extend linalg vectorization to MatmulOp.

The change is mostly renaming/clean-up. I only have minor comments, feel free to land after addressing.

Fri, Jan 17, 11:58 AM · Restricted Project
ftynse accepted D72942: [mlir][Linalg] Extend linalg vectorization to MatmulOp.
Fri, Jan 17, 11:58 AM · Restricted Project
ftynse added a comment to D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.

To clarify further, LLVM dialect in MLIR is not a new IR, it's essentially a representation of LLVM IR using MLIR. It's supposed to minimize the cost of translation from MLIR to LLVM IR. While it does diverge from LLVM IR proper, it does so for modeling reasons: there's no first-class constants in MLIR, neither are there phi nodes. Changing the name of an enumerand locally to the LLVM dialect will increase the cost of translation, which goes against the goal of this dialect. That being said, I do agree that we should use this opportunity to reflect on LLVM IR design and update it following the proper process.

Fri, Jan 17, 11:21 AM · Restricted Project
ftynse added a comment to D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.

Changing LLVM IR should _not_ be in this commit, it's semantically different and requires separate discussion with relevant people involved. I also don't see a compelling reason to block this from landing until LLVM IR changes (or decides not to). LLVM dialect models LLVM IR in its current state, whatever the state is.

Fri, Jan 17, 11:21 AM · Restricted Project
ftynse committed rGf343544b8138: [mlir] Generator converting LLVM intrinsics defs to MLIR ODS (authored by ftynse).
[mlir] Generator converting LLVM intrinsics defs to MLIR ODS
Fri, Jan 17, 9:25 AM
ftynse closed D72926: [mlir] Generator converting LLVM intrinsics defs to MLIR ODS.
Fri, Jan 17, 9:25 AM · Restricted Project
ftynse updated the diff for D72926: [mlir] Generator converting LLVM intrinsics defs to MLIR ODS.

address review

Fri, Jan 17, 9:25 AM · Restricted Project
ftynse added a comment to D72926: [mlir] Generator converting LLVM intrinsics defs to MLIR ODS.

I would like to make it a bit more useful (e.g. let the caller insert custom builders/printers/parsers) before plugging it into the flow directly. So it will come in several follow-ups. I'll make the requested changes directly in the commit.

Fri, Jan 17, 9:25 AM · Restricted Project
ftynse committed rGa922e23101b8: [mlir] Improve documentation in ModuleTranslation MLIR to LLVM IR (authored by ftynse).
[mlir] Improve documentation in ModuleTranslation MLIR to LLVM IR
Fri, Jan 17, 9:06 AM
ftynse added a comment to D72846: [mlir] support translation of multidimensional vectors to LLVM IR.

Added in a922e23101b882e22fc14ffb78b5857954dc86f3

Fri, Jan 17, 9:05 AM · Restricted Project
ftynse retitled D72542: [Flang][mlir] add a band-aid to support the creation of mutually recursive types when lowering to LLVM IR from [Flang] add a band-aid to support the creation of mutually recursive types when lowering to LLVM IR to [Flang][mlir] add a band-aid to support the creation of mutually recursive types when lowering to LLVM IR.
Fri, Jan 17, 8:55 AM · Restricted Project, Restricted Project
ftynse accepted D72542: [Flang][mlir] add a band-aid to support the creation of mutually recursive types when lowering to LLVM IR.
Fri, Jan 17, 8:55 AM · Restricted Project, Restricted Project
ftynse added a comment to D72924: [mlir][spirv] Add `const` qualifier for static arrays.

https://llvm.org/docs/DeveloperPolicy.html#code-reviews point 3

Fri, Jan 17, 8:55 AM · Restricted Project
ftynse added a comment to D72863: [mlir][Linalg] Add tensor support to Linalg EDSC Builders.

I'm not sure I understand what happens to the Op operands that are supposed to be tensors in individual linalg_ constructors. The comment says outputs should not contain tensors, but in individual constructors there's no generic outputs, and some StructuredIndexed *must* be passed in.

Fri, Jan 17, 8:49 AM · Restricted Project
ftynse added a comment to D72924: [mlir][spirv] Add `const` qualifier for static arrays.

I think you can push fixes for upstream breakages, especially in the part you own, without pre-review.

Fri, Jan 17, 8:26 AM · Restricted Project
ftynse accepted D72924: [mlir][spirv] Add `const` qualifier for static arrays.
Fri, Jan 17, 8:25 AM · Restricted Project
ftynse added reviewers for D72926: [mlir] Generator converting LLVM intrinsics defs to MLIR ODS: antiagainst, andydavis1, marcello.maggioni.
Fri, Jan 17, 7:55 AM · Restricted Project
ftynse created D72926: [mlir] Generator converting LLVM intrinsics defs to MLIR ODS.
Fri, Jan 17, 7:55 AM · Restricted Project

Thu, Jan 16

ftynse committed rGa4a42160c446: [mlir] support translation of multidimensional vectors to LLVM IR (authored by ftynse).
[mlir] support translation of multidimensional vectors to LLVM IR
Thu, Jan 16, 3:15 PM
ftynse closed D72846: [mlir] support translation of multidimensional vectors to LLVM IR.
Thu, Jan 16, 3:13 PM · Restricted Project
ftynse committed rG73f371c31d27: [mlir] NFC: Fix trivial typos (authored by kiszk).
[mlir] NFC: Fix trivial typos
Thu, Jan 16, 3:04 PM
ftynse closed D72672: [mlir] NFC: Fix trivial typos.
Thu, Jan 16, 3:04 PM · Restricted Project
ftynse accepted D72854: [mlir] Generalize m_ConstantIndexOp.
Thu, Jan 16, 8:57 AM · Restricted Project
ftynse added inline comments to D72854: [mlir] Generalize m_ConstantIndexOp.
Thu, Jan 16, 8:38 AM · Restricted Project
ftynse added inline comments to D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.
Thu, Jan 16, 8:28 AM · Restricted Project
ftynse added inline comments to D72834: [mlir] LLVM import: handle constant data and array/vector aggregates.
Thu, Jan 16, 7:30 AM · Restricted Project
ftynse added a reviewer for D72846: [mlir] support translation of multidimensional vectors to LLVM IR: aartbik.
Thu, Jan 16, 7:24 AM · Restricted Project
ftynse created D72846: [mlir] support translation of multidimensional vectors to LLVM IR.
Thu, Jan 16, 7:22 AM · Restricted Project
ftynse added a reviewer for D72839: [mlir] Harden error propagation in LLVM import: jmolloy.
Thu, Jan 16, 5:36 AM · Restricted Project
ftynse created D72839: [mlir] Harden error propagation in LLVM import.
Thu, Jan 16, 5:36 AM · Restricted Project
ftynse added a comment to rG0133cc60e4e2: Revert "[mlir] Create a gpu.module operation for the GPU Dialect.".

Could you please elaborate what exactly was causing problems? The original diff is https://reviews.llvm.org/D72336.

Thu, Jan 16, 4:30 AM
ftynse created D72834: [mlir] LLVM import: handle constant data and array/vector aggregates.
Thu, Jan 16, 4:27 AM · Restricted Project

Wed, Jan 15

ftynse added inline comments to D72542: [Flang][mlir] add a band-aid to support the creation of mutually recursive types when lowering to LLVM IR.
Wed, Jan 15, 2:27 PM · Restricted Project, Restricted Project
ftynse accepted D72766: [mlir][Linalg] NFC - Cleanup Linalg Pass locations and namespacing.
Wed, Jan 15, 7:07 AM · Restricted Project
ftynse accepted D72697: [mlir] fix broken link to Glossary.
Wed, Jan 15, 6:48 AM · Restricted Project
ftynse committed rGdc553ce646cd: [mlir] LLVM import: handle function-typed constants (authored by ftynse).
[mlir] LLVM import: handle function-typed constants
Wed, Jan 15, 6:02 AM
ftynse added a comment to D72129: [mlir] Add in-dialect lowering of gpu.all_reduce..

@mehdi_amini @nicolasvasilache @herhut Let's take some time and discuss builder APIs outside this diff (also involving @rriddle). My basic observations are that (1) writing structured IR, as in "with nested regions", looks unnecessarily complicated with builders, arguments are the same as those against goto-style programming; (2) a lot of IR construction internally happens in rewrite patterns, where location almost always remains the same, that of the matched operation root; (3) current EDSC APIs are contentious partly because it is unclear when reading the code when the function call creates the IR vs. when it's just a function call.

Wed, Jan 15, 4:29 AM · Restricted Project
ftynse added inline comments to D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.
Wed, Jan 15, 2:28 AM · Restricted Project
ftynse requested changes to D72741: [MLIR] LLVM dialect: Add llvm.atomicrmw.

Thanks! Most of my comments are minor.

Wed, Jan 15, 1:23 AM · Restricted Project
ftynse added a comment to D72585: [MLIR] Fixes for BUILD_SHARED_LIBS=on.

I think this goes into the right direction. The issue with this patch is that it removes some libraries from add_dependency for some targets, but not all of them, thus creating more inconsistency than we already have. Could you make it consistent for all targets and, if possible and in a follow up, add a doc that says how to properly set up a library for a new dialect.

Wed, Jan 15, 12:08 AM · Restricted Project

Tue, Jan 14

ftynse accepted D72481: [mlir][EDSC] Refactor dependencies involving EDSCs..
Tue, Jan 14, 11:58 PM · Restricted Project
ftynse added a comment to D72697: [mlir] fix broken link to Glossary.

The website uses a different layout anyway so many links won't work on GitHub, we don't expect them to. We prefer to be consistent and have the website working correctly.

Tue, Jan 14, 11:49 PM · Restricted Project
ftynse added a comment to D72129: [mlir] Add in-dialect lowering of gpu.all_reduce..

What is the status re EDSC discussion?
Were the pointers I sent offline enough to give a good picture / do you see how to followup on this to use (and maybe extend) EDSCs?

Tue, Jan 14, 1:59 PM · Restricted Project
ftynse accepted D72555: [mlir][Linalg] Update the semantics, verifier and test for Linalg with tensors..

Something went wrong with formatting in the commit description (mixed double space prefix and backticks?). Please reformat and feel free to land.

Tue, Jan 14, 1:56 PM · Restricted Project
ftynse requested changes to D72697: [mlir] fix broken link to Glossary.

It's in the website repo https://github.com/llvm/mlir-www/tree/master/website/content/getting_started and visible on mlir.llvm.org. Can you only fix the broken reference?

Tue, Jan 14, 7:43 AM · Restricted Project
ftynse requested changes to D72555: [mlir][Linalg] Update the semantics, verifier and test for Linalg with tensors..
Tue, Jan 14, 7:24 AM · Restricted Project
ftynse committed rGd6ea8ff0d74b: [mlir] Fix translation of splat constants to LLVM IR (authored by ftynse).
[mlir] Fix translation of splat constants to LLVM IR
Tue, Jan 14, 3:40 AM
ftynse closed D72688: [mlir] Fix translation of splat constants to LLVM IR.
Tue, Jan 14, 3:40 AM · Restricted Project
ftynse added inline comments to D72542: [Flang][mlir] add a band-aid to support the creation of mutually recursive types when lowering to LLVM IR.
Tue, Jan 14, 2:48 AM · Restricted Project, Restricted Project
ftynse added reviewers for D72688: [mlir] Fix translation of splat constants to LLVM IR: nicolasvasilache, aartbik.
Tue, Jan 14, 2:33 AM · Restricted Project
ftynse created D72688: [mlir] Fix translation of splat constants to LLVM IR.
Tue, Jan 14, 2:33 AM · Restricted Project

Mon, Jan 13

ftynse committed rG81e7922e83cf: [mlir] m_Constant() (authored by chelini).
[mlir] m_Constant()
Mon, Jan 13, 8:22 AM