Page MenuHomePhabricator

mravishankar (Mahesh Ravishankar)
User

Projects

User does not belong to any projects.

User Details

User Since
Jan 6 2015, 1:11 PM (429 w, 3 d)

Recent Activity

Tue, Mar 28

mravishankar accepted D147095: [mlir][bufferization] Use rewriter to erase ops in scf.forall bufferization..
Tue, Mar 28, 4:50 PM · Restricted Project, Restricted Project

Mon, Mar 20

mravishankar committed rG411b1d8f0795: [mlir][Tensor] Make `TilingInterface` implementation only return handle to the… (authored by mravishankar).
[mlir][Tensor] Make `TilingInterface` implementation only return handle to the…
Mon, Mar 20, 2:23 PM · Restricted Project, Restricted Project
mravishankar closed D146439: [mlir][Tensor] Make `TilingInterface` implementation only return handle to the created `pad` operation..
Mon, Mar 20, 2:23 PM · Restricted Project, Restricted Project
mravishankar committed rGd0e507f56990: [mlir][Tensor] Fix build error due to missing `<>` in D146440. (authored by mravishankar).
[mlir][Tensor] Fix build error due to missing `<>` in D146440.
Mon, Mar 20, 2:22 PM · Restricted Project, Restricted Project
mravishankar closed D146458: [mlir][Tensor] Fix build error due to missing `<>` in D146440..
Mon, Mar 20, 2:22 PM · Restricted Project, Restricted Project
mravishankar requested review of D146458: [mlir][Tensor] Fix build error due to missing `<>` in D146440..
Mon, Mar 20, 2:21 PM · Restricted Project, Restricted Project
mravishankar committed rGc21e88cc0261: [mlir][Tensor] Avoid dropping attributes for `tensor.pad` operations during… (authored by mravishankar).
[mlir][Tensor] Avoid dropping attributes for `tensor.pad` operations during…
Mon, Mar 20, 2:04 PM · Restricted Project, Restricted Project
mravishankar closed D146440: [mlir][Tensor] Avoid dropping attributes for `tensor.pad` operations during canonicalization..
Mon, Mar 20, 2:04 PM · Restricted Project, Restricted Project
mravishankar committed rG3af1c48c66a3: Changes to `SCFFuseProducerOfSliceResult` to also return the operations created… (authored by mravishankar).
Changes to `SCFFuseProducerOfSliceResult` to also return the operations created…
Mon, Mar 20, 1:56 PM · Restricted Project, Restricted Project
mravishankar closed D146254: Changes to `SCFFuseProducerOfSliceResult` to also return the operations created during fusion..
Mon, Mar 20, 1:56 PM · Restricted Project, Restricted Project
mravishankar updated the diff for D146254: Changes to `SCFFuseProducerOfSliceResult` to also return the operations created during fusion..

Rebase

Mon, Mar 20, 11:59 AM · Restricted Project, Restricted Project
mravishankar added a reviewer for D146440: [mlir][Tensor] Avoid dropping attributes for `tensor.pad` operations during canonicalization.: hanchung.
Mon, Mar 20, 11:31 AM · Restricted Project, Restricted Project
mravishankar added a reviewer for D146439: [mlir][Tensor] Make `TilingInterface` implementation only return handle to the created `pad` operation.: hanchung.
Mon, Mar 20, 11:30 AM · Restricted Project, Restricted Project
mravishankar added a reviewer for D146254: Changes to `SCFFuseProducerOfSliceResult` to also return the operations created during fusion.: hanchung.
Mon, Mar 20, 11:30 AM · Restricted Project, Restricted Project
mravishankar requested review of D146440: [mlir][Tensor] Avoid dropping attributes for `tensor.pad` operations during canonicalization..
Mon, Mar 20, 10:13 AM · Restricted Project, Restricted Project
mravishankar requested review of D146439: [mlir][Tensor] Make `TilingInterface` implementation only return handle to the created `pad` operation..
Mon, Mar 20, 10:13 AM · Restricted Project, Restricted Project

Thu, Mar 16

mravishankar requested review of D146254: Changes to `SCFFuseProducerOfSliceResult` to also return the operations created during fusion..
Thu, Mar 16, 12:42 PM · Restricted Project, Restricted Project
mravishankar committed rG1da04b04cd59: Fix asan error uncovered after landing D145133. (authored by mravishankar).
Fix asan error uncovered after landing D145133.
Thu, Mar 16, 11:06 AM · Restricted Project, Restricted Project
mravishankar closed D146249: Fix asan error uncovered after landing D145133..
Thu, Mar 16, 11:06 AM · Restricted Project, Restricted Project
mravishankar requested review of D146249: Fix asan error uncovered after landing D145133..
Thu, Mar 16, 11:01 AM · Restricted Project, Restricted Project
mravishankar committed rG809e3d8c98a8: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the… (authored by mravishankar).
[mlir][TilingInterface] Modify `TilingInterface` methods to better return the…
Thu, Mar 16, 7:29 AM · Restricted Project, Restricted Project
mravishankar closed D145133: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR..
Thu, Mar 16, 7:29 AM · Restricted Project, Restricted Project
mravishankar updated the diff for D145133: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR..

Rebase.

Thu, Mar 16, 6:36 AM · Restricted Project, Restricted Project

Tue, Mar 14

mravishankar committed rGd3fa067e699c: [mlir][Tensor] Use folded evaluators in tiling implementation of `tensor.pad`. (authored by mravishankar).
[mlir][Tensor] Use folded evaluators in tiling implementation of `tensor.pad`.
Tue, Mar 14, 8:29 PM · Restricted Project, Restricted Project
mravishankar closed D145135: [mlir][Tensor] Use folded evaluators in tiling implementation of `tensor.pad`..
Tue, Mar 14, 8:29 PM · Restricted Project, Restricted Project

Mon, Mar 13

mravishankar added a reviewer for D145135: [mlir][Tensor] Use folded evaluators in tiling implementation of `tensor.pad`.: springerm.
Mon, Mar 13, 10:24 AM · Restricted Project, Restricted Project
mravishankar updated the diff for D145133: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR..

Rebase again.

Mon, Mar 13, 10:12 AM · Restricted Project, Restricted Project
mravishankar updated the diff for D145133: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR..

Rebase

Mon, Mar 13, 10:02 AM · Restricted Project, Restricted Project

Fri, Mar 3

mravishankar accepted D145250: [mlir][Interfaces] ReifyRankedShapedTypeOpInterface returns OpFoldResults.

Awesome! Thanks!

Fri, Mar 3, 10:35 AM · Restricted Project, Restricted Project

Mar 1 2023

mravishankar updated the diff for D145133: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR..

Fix lit tests.

Mar 1 2023, 10:41 PM · Restricted Project, Restricted Project
mravishankar updated the diff for D145135: [mlir][Tensor] Use folded evaluators in tiling implementation of `tensor.pad`..

Fix bug.

Mar 1 2023, 8:30 PM · Restricted Project, Restricted Project
mravishankar requested review of D145135: [mlir][Tensor] Use folded evaluators in tiling implementation of `tensor.pad`..
Mar 1 2023, 6:23 PM · Restricted Project, Restricted Project
mravishankar added reviewers for D145133: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR.: pifon2a, ftynse, hanchung, springerm.
Mar 1 2023, 5:11 PM · Restricted Project, Restricted Project
mravishankar requested review of D145133: [mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR..
Mar 1 2023, 5:10 PM · Restricted Project, Restricted Project

Feb 17 2023

mravishankar resigned from D144108: [mlir][LinAlg][Transform] Add a transform op for conv2d to im2col.

I am unblocking given it will be iterated on. Happy to chat offline later

Feb 17 2023, 7:38 PM · Restricted Project, Restricted Project
mravishankar requested changes to D144294: [mlir][linalg] Add pattern to convert batch_matmul to matmul.

Probably needs to be into NamedOpConversion pass

Feb 17 2023, 7:36 PM · Restricted Project, Restricted Project

Feb 16 2023

mravishankar accepted D144151: [mlir][tensor] Do not use affine ops on non-perfect-tiling unpack cases..

Maybe we should file a bug to fix the issue with simplification

Feb 16 2023, 11:28 PM · Restricted Project, Restricted Project
mravishankar requested changes to D144067: [mlir][linalg] Add more shape information for tensor.pack generalization.

This seems strange to me. There is nothing that prevents a subsequent transformation to collapse the cast into the slice. Indeed it is probably better to do so.

Feb 16 2023, 11:20 PM · Restricted Project, Restricted Project
mravishankar accepted D143655: [mlir][linalg] Support non-function ops in Detensorize.

Thanks!

Feb 16 2023, 9:56 PM · Restricted Project, Restricted Project

Feb 15 2023

mravishankar requested changes to D144108: [mlir][LinAlg][Transform] Add a transform op for conv2d to im2col.

I know this pattern exists in IREE, but it is a bit of a hack. The representation of the im2col as a linalg.generic doesnt work as well. In reality it is similar to a gather. If this is upstreamed, it might be worth doing this right and not use an unnecessarily higher-dimensional operation for representing the im2col. Maybe we can chat offline?

Feb 15 2023, 10:55 AM · Restricted Project, Restricted Project

Feb 12 2023

mravishankar accepted D143839: Add support for processing Ops with empty region to the CSE.
Feb 12 2023, 9:32 AM · Restricted Project, Restricted Project

Feb 11 2023

mravishankar added a comment to D141900: Add support for processing Ops with empty region to the CSE.

Sure. Just push the test. Thanks!

Feb 11 2023, 8:36 AM · Restricted Project, Restricted Project

Feb 9 2023

mravishankar added a comment to D141900: Add support for processing Ops with empty region to the CSE.

@Ding I tried to rebase this and land. Seems to fail tests.

Feb 9 2023, 10:19 PM · Restricted Project, Restricted Project
mravishankar added a comment to D141900: Add support for processing Ops with empty region to the CSE.

I don't think so. I am AFK. I can merge when I get to my machine

Feb 9 2023, 8:51 PM · Restricted Project, Restricted Project
mravishankar accepted D143686: [mlir][tensor] Add shape inference methods to tensor::PackOp..
Feb 9 2023, 4:45 PM · Restricted Project, Restricted Project
mravishankar requested changes to D143655: [mlir][linalg] Support non-function ops in Detensorize.

I see why you made it a function pass, but I think it needs to be a pass on FunctionOpInterface. There are examples of this in IREE https://github.com/iree-org/iree/blob/55d0ffe7ea4ca5b58dd9197608151cc2f8c7521c/compiler/src/iree/compiler/Dialect/Flow/Transforms/Passes.td#L65 .

Feb 9 2023, 3:32 PM · Restricted Project, Restricted Project
mravishankar accepted D142130: [mlir][bufferization] Support multiple leaves in EmptyTensorElimination.

Have one comment here, more for my understanding of the nuances. Change itself looks good!

Feb 9 2023, 10:03 AM · Restricted Project, Restricted Project

Jan 31 2023

mravishankar committed rG01890942f0c0: Fix missing return of LogicalResult from D142848 (authored by mravishankar).
Fix missing return of LogicalResult from D142848
Jan 31 2023, 6:07 PM · Restricted Project, Restricted Project
mravishankar closed D143023: Fix missing return of LogicalResult from D142848.
Jan 31 2023, 6:07 PM · Restricted Project, Restricted Project
mravishankar abandoned D143011: Fix a missing return-on-failure missed with D142848..
Jan 31 2023, 3:22 PM · Restricted Project, Restricted Project
mravishankar requested review of D143023: Fix missing return of LogicalResult from D142848.
Jan 31 2023, 3:21 PM · Restricted Project, Restricted Project
mravishankar added a reviewer for D143011: Fix a missing return-on-failure missed with D142848.: hanchung.
Jan 31 2023, 1:30 PM · Restricted Project, Restricted Project
mravishankar requested review of D143011: Fix a missing return-on-failure missed with D142848..
Jan 31 2023, 1:29 PM · Restricted Project, Restricted Project
mravishankar committed rG69011a2ad0ce: [mlir][Linalg] Make Elementwise op fusion return a map from existing values to… (authored by mravishankar).
[mlir][Linalg] Make Elementwise op fusion return a map from existing values to…
Jan 31 2023, 12:34 PM · Restricted Project, Restricted Project
mravishankar closed D142848: [mlir][Linalg] Make Elementwise op fusion return a map from existing values to values in the fused op..
Jan 31 2023, 12:34 PM · Restricted Project, Restricted Project

Jan 30 2023

mravishankar added inline comments to D142848: [mlir][Linalg] Make Elementwise op fusion return a map from existing values to values in the fused op..
Jan 30 2023, 3:26 PM · Restricted Project, Restricted Project
mravishankar updated the diff for D142848: [mlir][Linalg] Make Elementwise op fusion return a map from existing values to values in the fused op..

Address comments.

Jan 30 2023, 3:26 PM · Restricted Project, Restricted Project

Jan 29 2023

mravishankar requested review of D142848: [mlir][Linalg] Make Elementwise op fusion return a map from existing values to values in the fused op..
Jan 29 2023, 1:59 PM · Restricted Project, Restricted Project

Jan 26 2023

mravishankar added a comment to D142480: [mlir][SCF] CSE and hoist operations in scf.if branches.

If a tensor.insert_slice can be hoisted, bufferization should be immune to that (I'd think it would even be better, but dont know the example that you have in mind). Hoisting entire then and else branches and ending up with a select seems to be a good thing in general.

Yes it is actually better for the bufferization — a flat structure is easier to bufferize than one with nested blocks; fewer things can go wrong.

But hoisting things also means that ops that were previously inside a branch and executed 50% of the time are now executed 100% of the time. (After bufferization, pure tensor ops have a side effect and therefore a cost.)

Jan 26 2023, 3:10 PM · Restricted Project, Restricted Project
mravishankar added a comment to D142557: [mlir][Linalg] Fix crash when converting rank-reduced tensor.insert_slice.

I think the core logic being fixed here is not something I worked on, so I dont have full context. Adding folks who worked on this as reviewers...

Jan 26 2023, 9:32 AM · Restricted Project, Restricted Project
mravishankar added reviewers for D142557: [mlir][Linalg] Fix crash when converting rank-reduced tensor.insert_slice: pifon2a, herhut.
Jan 26 2023, 9:31 AM · Restricted Project, Restricted Project
mravishankar added inline comments to D142557: [mlir][Linalg] Fix crash when converting rank-reduced tensor.insert_slice.
Jan 26 2023, 9:30 AM · Restricted Project, Restricted Project

Jan 25 2023

mravishankar added a comment to D142558: [mlir][transforms] Simplify OperationEquivalence and CSE.

This is pretty cool! It makes CSE very powerful. Last time I edited, I tried to walk a fine line and take small steps. So thanks for cleaning all of these up! This is taking a big step. I think its OK. But I'll probably wait for others to chime in.

Jan 25 2023, 10:17 PM · Restricted Project, Restricted Project
mravishankar added a comment to D142480: [mlir][SCF] CSE and hoist operations in scf.if branches.

Then I am confused. If it is speculatable only, then you can just hoist.. Then CSE will handle the equivalence and replacing uses.... Not sure what I am missing....

What I meant:

%r = scf.if ... {
  "non_specutable_op"()
  ...
} else {
  "non_specutable_op"()
  ...
}

==> /* This transformation does not affect performance. */

"non_specutable_op"()
%r = scf.if ... {
  ...
} else {
  ...
}

Hoisting and CSE'ing "non_spectuable_op" from both branches is OK. Any op for that matter. Just hoisting from the "then" branch is not OK; that's only allowed if the op is pure and specutable.

My main concerns about "just hoisting" are performance implications. Should we hoist a tensor.insert_slice from an scf.if? That's a buffer copy after bufferization, so it can lead to a performance regression. We could go even further: Hoist the entire "then" and "else" bodies (assuming they are all pure+specutable ops; usually the case in tensor IR), so that only the terminators are left over; scf.if will become an arith.select. That's probably not desirable.

Jan 25 2023, 10:07 PM · Restricted Project, Restricted Project

Jan 24 2023

mravishankar accepted D142206: [mlir][linalg] Convert tensor.from_elements to destination style.
Jan 24 2023, 10:20 AM · Restricted Project, Restricted Project
mravishankar added a comment to D142480: [mlir][SCF] CSE and hoist operations in scf.if branches.

You are CSE-ing and hoisting... thats already within the preview of treating it as a speculatable op. Doing this on non-speculatable ops would be illegal AFAIK.

That’s why this pass is specific to scf.if. It does not hoist or CSE non-specutable ops from other ops. That should be legal.

Jan 24 2023, 10:20 AM · Restricted Project, Restricted Project
mravishankar added a comment to D142480: [mlir][SCF] CSE and hoist operations in scf.if branches.

Could we just use the hoisting part here. You can just hoist operations outside of if and let standard CSE common out the users.... THis is duplicating a lot of functionality

We could but this allows us to CSE irrespective of op spectuability and it has no effect on performance (because CSE’d ops are executed in either branch).

Jan 24 2023, 10:03 AM · Restricted Project, Restricted Project
mravishankar requested changes to D142480: [mlir][SCF] CSE and hoist operations in scf.if branches.

Could we just use the hoisting part here. You can just hoist operations outside of if and let standard CSE common out the users.... THis is duplicating a lot of functionality

Jan 24 2023, 9:25 AM · Restricted Project, Restricted Project

Jan 23 2023

mravishankar added a comment to D142206: [mlir][linalg] Convert tensor.from_elements to destination style.

This is very painful... Maybe we can just hold off on this until it is needed....

Jan 23 2023, 12:28 PM · Restricted Project, Restricted Project
mravishankar accepted D142207: [mlir][linalg] Convert tensor.pad to destination style.

Unblocking, but have one comment remaining.

Jan 23 2023, 12:27 PM · Restricted Project, Restricted Project
mravishankar accepted D142205: [mlir][linalg] Convert tensor.generate to destination style.
Jan 23 2023, 12:24 PM · Restricted Project, Restricted Project
mravishankar accepted D141900: Add support for processing Ops with empty region to the CSE.

Thanks!

Jan 23 2023, 12:23 PM · Restricted Project, Restricted Project

Jan 20 2023

mravishankar added a comment to D142205: [mlir][linalg] Convert tensor.generate to destination style.

The pattern itself looks fine, but I think this has a dependency issue. There is already a Tensor -> Linalg dependence. This is adding a Linalg -> Tensor dependence.... That could be problematic... is there a way to do this as part of bufferization or something, or move these patterns to Linalg?

Note it is a dependency from TensorTransforms -> Linalg and we already have it without this change. The implementation is actually largely copied from Tensor/Transforms/BufferizableOpInterfaceImpl.cpp.

Jan 20 2023, 2:51 PM · Restricted Project, Restricted Project
mravishankar requested changes to D142205: [mlir][linalg] Convert tensor.generate to destination style.

The pattern itself looks fine, but I think this has a dependency issue. There is already a Tensor -> Linalg dependence. This is adding a Linalg -> Tensor dependence.... That could be problematic... is there a way to do this as part of bufferization or something, or move these patterns to Linalg?

Jan 20 2023, 11:28 AM · Restricted Project, Restricted Project
mravishankar accepted D141807: [mlir] Remove Linalg fusion-on-memrefs..

unblocking since there was a PSA. Thanks!

Jan 20 2023, 9:42 AM · Restricted Project, Restricted Project
mravishankar requested changes to D141900: Add support for processing Ops with empty region to the CSE.

Looks good. THanks. Just a few nits.

Jan 20 2023, 9:42 AM · Restricted Project, Restricted Project
mravishankar added inline comments to D141900: Add support for processing Ops with empty region to the CSE.
Jan 20 2023, 9:40 AM · Restricted Project, Restricted Project
mravishankar accepted D141977: [mlir][tensor] Add a tiling interface registration for pack/unpack only..

Yeah, this is fine by me.

Jan 20 2023, 9:39 AM · Restricted Project, Restricted Project
mravishankar requested changes to D142207: [mlir][linalg] Convert tensor.pad to destination style.
Jan 20 2023, 9:38 AM · Restricted Project, Restricted Project

Jan 19 2023

mravishankar added a comment to D141804: [mlir][linalg] Omit printing result types for named ops..

For instance developers often store models/IR in linalg format to be able to work in isolation of the front end. Those kind of changes forces them to regenerate models.

I think this is the only valid argument against landing this PR in this whole thread.

Jan 19 2023, 9:34 PM · Restricted Project, Restricted Project

Jan 17 2023

mravishankar added a comment to D141804: [mlir][linalg] Omit printing result types for named ops..

To start with, I see at minima design consistency with the general dialect design I think I saw upstream so far.

Jan 17 2023, 12:01 PM · Restricted Project, Restricted Project
mravishankar requested changes to D141900: Add support for processing Ops with empty region to the CSE.

Thanks! This was the fix I had in mind.

Jan 17 2023, 11:14 AM · Restricted Project, Restricted Project
mravishankar added a comment to D141804: [mlir][linalg] Omit printing result types for named ops..

From my experience integrating MLIR into Google's internal repository, breakages from changes like these seem to be very frequent (e.g., the two already mentioned above), so I don't see why this change should be blocked while others get a pass. If we want to be more careful about API breakages, then we should first formalize this somehow and apply the same rules to everyone.
I never considered the printed version of MLIR stable, and I'm not sure that its stability should be the goal that prevents improvements to the printing format.
Regarding churn, we should not forget that not improving the printed format also causes churn every time someone is reading or modifying it.
I think that redundancy in IR causes more issues than it solves, so I would prefer landing this change.

This is an unnecessary change that just causes downstream churn. Not sure it is worth it

Looking at the previous discussion, I think that people's opinions differ on this one. However, I would like to touch on the phrasing of this comment. The very fact that someone spent the time to make this change means that they do think it is necessary. Thus, in my opinion, stating that it is unnecessary in such absolutist terms is disrespectful. The same point could have been achieved by saying something like "I disagree with this change because <your argumentation>". Even if you know Alex personally, and you're sure he wouldn't mind the statement, this is still a public forum where others (maybe potential future contributors) will read these discussions. I think we should be more careful in how we communicate, and make sure that this forum is an inclusive place where others' opinions are respected.

Jan 17 2023, 11:04 AM · Restricted Project, Restricted Project
mravishankar added a comment to D141804: [mlir][linalg] Omit printing result types for named ops..

As a meta-point: I'm concerned about the push-back on improving things based on churn. The comment about users wanting stability can only be an argument about spending more time on design and being more thorough before adding anything to the codebase. However that means much less velocity and higher cost to experiment. So far this isn't the tradeoff we've been making in MLIR I believe.

Jan 17 2023, 11:02 AM · Restricted Project, Restricted Project

Jan 16 2023

mravishankar added a comment to D141804: [mlir][linalg] Omit printing result types for named ops..

I am in favor of landing this (but please ensure consensus before doing so). I have argued against changes before that were purely aesthetic but introduced a lot of churn (like the one in https://reviews.llvm.org/D133076). Ultimately it is a subjective decision. I have a different opinion in this case because the linalg dialect has a smaller usage scope compared to arith, so I am willing to accept more churn. Also, this change improves ergonomics when reading linalg IR, which we increasingly do and hence care about. This was also one of the motivations to introduce new operations to linalg like map.

Jan 16 2023, 5:37 PM · Restricted Project, Restricted Project
mravishankar added a comment to D141028: [mlir][TilingInterface] Add an option to tile and fuse to yield replacement for the fused producer..

This patch had a use-after-free issue. Fixed with https://reviews.llvm.org/D141869

Jan 16 2023, 2:02 PM · Restricted Project, Restricted Project
mravishankar added a comment to D141804: [mlir][linalg] Omit printing result types for named ops..

I think this is worth cleaning this up. Small things like these can make it look like linalg is a second-class dialect where we don't care about code quality.

Jan 16 2023, 1:12 PM · Restricted Project, Restricted Project
mravishankar committed rGdbbd9070152a: [mlir][TilingInterface] Fix use after free error from D141028. (authored by mravishankar).
[mlir][TilingInterface] Fix use after free error from D141028.
Jan 16 2023, 1:00 PM · Restricted Project, Restricted Project
mravishankar closed D141869: [mlir][TilingInterface] Fix use after free error from D141028..
Jan 16 2023, 1:00 PM · Restricted Project, Restricted Project
mravishankar requested review of D141869: [mlir][TilingInterface] Fix use after free error from D141028..
Jan 16 2023, 12:52 PM · Restricted Project, Restricted Project
mravishankar abandoned D134307: [mlir][TilingInterface] Add callback to yield a produced value..
Jan 16 2023, 10:38 AM · Restricted Project, Restricted Project
mravishankar committed rG9db7d4edd82e: [mlir][TilingInterface] Add an option to tile and fuse to yield replacement for… (authored by mravishankar).
[mlir][TilingInterface] Add an option to tile and fuse to yield replacement for…
Jan 16 2023, 10:36 AM · Restricted Project, Restricted Project
mravishankar closed D141028: [mlir][TilingInterface] Add an option to tile and fuse to yield replacement for the fused producer..
Jan 16 2023, 10:36 AM · Restricted Project, Restricted Project
mravishankar added a comment to D141804: [mlir][linalg] Omit printing result types for named ops..

This is an unnecessary change that just causes downstream churn. Not sure it is worth it

It makes IR shorter and nicer. It is not a lot of churn, the named ops are not as widely used as linalg.generic. 90% of the tests were fixed by a simple sed command with s/ outs(\(.*\)) -> .*$/ outs(\1)/g.

The "churn" aspect is subjective. There were patches like https://github.com/llvm/llvm-project/commit/abc362a1077b9cb4186e3e53a616589c7fed4387, that were not blocked.

Jan 16 2023, 10:33 AM · Restricted Project, Restricted Project

Jan 15 2023

mravishankar added inline comments to D141028: [mlir][TilingInterface] Add an option to tile and fuse to yield replacement for the fused producer..
Jan 15 2023, 10:19 PM · Restricted Project, Restricted Project
mravishankar updated the diff for D141028: [mlir][TilingInterface] Add an option to tile and fuse to yield replacement for the fused producer..

Address comments.

Jan 15 2023, 10:18 PM · Restricted Project, Restricted Project
mravishankar committed rGce349ff1a483: [mlir][TilingInterface] NFC: Separate out a utility method to perform one step… (authored by mravishankar).
[mlir][TilingInterface] NFC: Separate out a utility method to perform one step…
Jan 15 2023, 9:21 PM · Restricted Project, Restricted Project
mravishankar closed D141027: [mlir][TilingInterface] NFC: Separate out a utility method to perform one step of tile + fuse..
Jan 15 2023, 9:21 PM · Restricted Project, Restricted Project
mravishankar committed rG94f2a6ddde5c: [mlir][TilingInterface] NFC: Consolidate yield handling. (authored by mravishankar).
[mlir][TilingInterface] NFC: Consolidate yield handling.
Jan 15 2023, 9:19 PM · Restricted Project, Restricted Project
mravishankar closed D139392: [mlir][TilingInterface] NFC: Consolidate yield handling..
Jan 15 2023, 9:19 PM · Restricted Project, Restricted Project