Page MenuHomePhabricator

bondhugula (Uday Bondhugula)
User

Projects

User does not belong to any projects.

User Details

User Since
Dec 14 2019, 8:54 PM (56 w, 6 d)

Recent Activity

Wed, Jan 6

bondhugula requested changes to D94116: Update the maximum integer bitwidth in MLIR. .

Can you please add a proper commit summary for the rationale (basically whatever is in the discussion now)? Also, as @jpienaar points out, the commit title isn't really accurate.

Wed, Jan 6, 6:14 PM · Restricted Project

Fri, Jan 1

bondhugula added a comment to D93947: [mlir] Fix indexing of first offset operand in ops that implement OffsetSizeAndStrideOpInterface.

Missing a test case?

Fri, Jan 1, 5:36 PM · Restricted Project
bondhugula committed rG3bcca6b12d4c: [MLIR] Fix affine_map compose with multi-symbols (authored by yaochengji).
[MLIR] Fix affine_map compose with multi-symbols
Fri, Jan 1, 5:31 PM
bondhugula closed D93831: fix affine_map compose with multi-symbols.
Fri, Jan 1, 5:31 PM · Restricted Project

Thu, Dec 31

bondhugula added a comment to D93831: fix affine_map compose with multi-symbols.

Thanks very much for fixing this!

Thu, Dec 31, 7:36 AM · Restricted Project

Thu, Dec 24

bondhugula requested changes to D93704: [mlir][Linalg] Introduce linalg.padded_view op..
Thu, Dec 24, 10:30 AM · Restricted Project

Wed, Dec 23

bondhugula added a comment to D93120: [mlir][vector] Add more vector Ops canonicalization.

@ThomasRaoux, Not on a directly related note, but this file VectorOps.cpp is taking nearly 12-15s to build on a fast workstation -- directly increasing the build times / critical path towards mlir-opt and other targets for everyone. It'd be great to consider refactoring among next steps.

Wed, Dec 23, 11:20 PM · Restricted Project
bondhugula added inline comments to D93717: [IR] Add an ImplicitLocOpBuilder helper class for building IR with the same loc..
Wed, Dec 23, 10:12 AM · Restricted Project
bondhugula added inline comments to D93717: [IR] Add an ImplicitLocOpBuilder helper class for building IR with the same loc..
Wed, Dec 23, 10:12 AM · Restricted Project

Tue, Dec 22

bondhugula committed rG3e07b0b9d336: [MLIR] Fix lowering of affine operations with return values (authored by pr4tgpt).
[MLIR] Fix lowering of affine operations with return values
Tue, Dec 22, 8:16 AM
bondhugula closed D93090: [MLIR] Fix lowering of affine operations with return values..
Tue, Dec 22, 8:16 AM · Restricted Project
bondhugula accepted D93090: [MLIR] Fix lowering of affine operations with return values..
Tue, Dec 22, 3:36 AM · Restricted Project
bondhugula accepted D93090: [MLIR] Fix lowering of affine operations with return values..
Tue, Dec 22, 3:11 AM · Restricted Project

Mon, Dec 21

bondhugula resigned from D93500: [mlir[[vector] Extend Transfer read/write ops to support tensor types..
Mon, Dec 21, 7:39 AM · Restricted Project

Sat, Dec 19

bondhugula requested changes to D93500: [mlir[[vector] Extend Transfer read/write ops to support tensor types..
Sat, Dec 19, 10:17 AM · Restricted Project
bondhugula requested changes to D93090: [MLIR] Fix lowering of affine operations with return values..

Mostly minor comments.

Sat, Dec 19, 12:47 AM · Restricted Project

Dec 11 2020

bondhugula added inline comments to D93090: [MLIR] Fix lowering of affine operations with return values..
Dec 11 2020, 12:14 AM · Restricted Project
bondhugula requested changes to D93090: [MLIR] Fix lowering of affine operations with return values..
Dec 11 2020, 12:11 AM · Restricted Project
bondhugula added a reviewer for D93090: [MLIR] Fix lowering of affine operations with return values.: dcaballe.
Dec 11 2020, 12:08 AM · Restricted Project

Dec 10 2020

bondhugula added inline comments to D92876: [mlir][Affine] Add support for multi-store producer fusion.
Dec 10 2020, 1:03 AM · Restricted Project
bondhugula requested changes to D92876: [mlir][Affine] Add support for multi-store producer fusion.

Thanks for this significant update! Some superficial comments to start with. It may be good to handle the pass documentation update as well, starting with an update to the summary. It's currently emptly: https://mlir.llvm.org/docs/Passes/#-affine-loop-fusion-fuse-affine-loop-nests
You could mention about the fusion strategies producer/consumer and sibling ones as well. Here's the previous doc paragraph that disappeared (was overwritten) when pass documentation was migrated to auto-generated stuff.

Dec 10 2020, 12:55 AM · Restricted Project
bondhugula added inline comments to D92770: [mlir] Support inlining into affine operations.
Dec 10 2020, 12:12 AM · Restricted Project
bondhugula added inline comments to D92770: [mlir] Support inlining into affine operations.
Dec 10 2020, 12:11 AM · Restricted Project
bondhugula added a comment to D92770: [mlir] Support inlining into affine operations.

Looks great!

Dec 10 2020, 12:03 AM · Restricted Project

Dec 9 2020

bondhugula added a comment to D92991: [mlir] Start splitting the `tensor` dialect out of `std`..

Thanks!

I'm -1 on the proposed name change: LLVM has {insert|extract}{element|value}, op names are globally consistent with that terminology.

Dec 9 2020, 11:56 PM · Restricted Project
bondhugula added a comment to D92991: [mlir] Start splitting the `tensor` dialect out of `std`..

Nothing major, but I tend to prefer tensor.get_element to tensor.extract for two reasons: (1) extract in standard data structures/ADT literature also involves removing the element out of the structure (eg. extract_min/extract_max for heaps, priority queues), (2) extract is more suitable for containers but this is an immutable entity.

Dec 9 2020, 10:26 PM · Restricted Project

Dec 7 2020

bondhugula committed rGdc930e5f2f91: [MLIR][Affine] Add affine.for normalization support (authored by navdeepkk).
[MLIR][Affine] Add affine.for normalization support
Dec 7 2020, 8:39 AM
bondhugula closed D92233: [MLIR][Affine] Add affine.for normalization support.
Dec 7 2020, 8:38 AM · Restricted Project
bondhugula accepted D92233: [MLIR][Affine] Add affine.for normalization support.
Dec 7 2020, 1:26 AM · Restricted Project

Dec 4 2020

bondhugula added inline comments to D92635: [mlir][vector][NFC] Change UnrollVectorPattern to not be statically dependent on an op type.
Dec 4 2020, 12:41 AM · Restricted Project
bondhugula added inline comments to D92614: [mlir][vector][avx512] move avx512 lowering pass into general vector lowering.
Dec 4 2020, 12:30 AM · Restricted Project

Dec 3 2020

bondhugula committed rGb2391d5f0da2: [MLIR] Normalize the results of normalizable operations (authored by imaihal).
[MLIR] Normalize the results of normalizable operations
Dec 3 2020, 6:05 AM
bondhugula closed D88719: [mlir] Normalize the results of normalizable operations.
Dec 3 2020, 6:05 AM · Restricted Project
bondhugula accepted D88719: [mlir] Normalize the results of normalizable operations.
Dec 3 2020, 5:50 AM · Restricted Project
bondhugula accepted D92233: [MLIR][Affine] Add affine.for normalization support.

This looks great! Thanks.

Dec 3 2020, 5:49 AM · Restricted Project

Dec 2 2020

bondhugula committed rGb276bf5a572b: [MLIR][NFC] Fix mix up between dialect attribute values and names (authored by bondhugula).
[MLIR][NFC] Fix mix up between dialect attribute values and names
Dec 2 2020, 1:05 PM
bondhugula closed D92502: [MLIR][NFC] Fix mix up between dialect attribute values and names.
Dec 2 2020, 1:05 PM · Restricted Project
bondhugula requested review of D92502: [MLIR][NFC] Fix mix up between dialect attribute values and names.
Dec 2 2020, 12:45 PM · Restricted Project

Nov 24 2020

bondhugula accepted D86380: [MLIR][Affine] Add custom builders for AffineVectorLoadOp/AffineVectorStoreOp.
Nov 24 2020, 11:13 AM · Restricted Project
bondhugula added inline comments to D91967: [mlir][std] Mark tensor_to_memref as no side-effect.
Nov 24 2020, 1:15 AM · Restricted Project

Nov 23 2020

bondhugula added a comment to D91982: [mlir] Add conversion from SCF parallel loops to OpenMP.

Nice. Looks like you missed adding test cases.

Nov 23 2020, 11:06 AM · Restricted Project
bondhugula added inline comments to D91967: [mlir][std] Mark tensor_to_memref as no side-effect.
Nov 23 2020, 10:26 AM · Restricted Project
bondhugula added inline comments to D91967: [mlir][std] Mark tensor_to_memref as no side-effect.
Nov 23 2020, 10:17 AM · Restricted Project

Nov 22 2020

bondhugula added a comment to D86380: [MLIR][Affine] Add custom builders for AffineVectorLoadOp/AffineVectorStoreOp.

It's been exactly two months since I submitted a review!! - this was I think a simple patch. I've forgotten all context! @dcaballe - could you please review this as it's directly on top of your work? @flaub please go with @dcaballe's review.

Nov 22 2020, 8:00 AM · Restricted Project

Nov 21 2020

bondhugula requested changes to D91863: [mlir][Linalg] NFC - Factor out Linalg functionality for shape and loop bounds computation.
Nov 21 2020, 8:20 PM · Restricted Project

Nov 18 2020

bondhugula added a comment to D91571: [mlir][IR] Use tablegen for the BuiltinDialect and operations.

I think the builtin dialect isn't documented anywhere:
https://mlir.llvm.org/docs/Dialects/
(Crazy question: do we really need to even call it a dialect? Can't it just be BuiltinOps.h?)

Nov 18 2020, 3:20 PM · Restricted Project

Nov 16 2020

bondhugula added a comment to D90899: [mlir] Add DecomposeCallGraphTypes pass..

Is "call graph types" the best term here? This doesn't have to do with call graphs but with function signatures.(?) The older naming along the lines of FunctionSignatureConversion or DecomposeFuncSignature or UnpackFuncSignature looks clearer than anything with "CallGraph".

I'm using the term "call graph types" as a stand-in for "function signatures, call signatures, and return signatures". That is, types that flow along call graph edges.

I was also keeping an eye towards eventually generalizing this beyond std.func/call/return. In that more general setting, I think "call graph" is much clearer, because different dialects might have objects in their call graphs that are not called "function". For example, the concept of a first class callable object has many different names: lambdas, closures, delegates. Using the name "function" doesn't sit well with me in that more general setting. In all cases though, the set of patterns we need are the same: one on CallOpInterface, one on CallableOpInterface, and one on something that represents returning control flow from a CallableOpInterface. Once we have extended those interfaces appropriately, the patterns in this pass would be updated to MatchAnyOpTypeTag patterns written over those interfaces and everything would just work.

I guess it depends on how one looks at the pass: is the pass fundamentally updating function signatures / callable signatures (and everything else is just changing to match), or is the pass to be viewed as decomposing types along call graph edges. I was thinking about the pass as the latter. What do you think?

Nov 16 2020, 1:03 PM · Restricted Project
bondhugula accepted D90899: [mlir] Add DecomposeCallGraphTypes pass..

LGTM - thanks! Just some minor comments (please see if the "mixed" test case could be improved for coverage).

Nov 16 2020, 9:40 AM · Restricted Project

Nov 12 2020

bondhugula added a comment to D91374: [MLIR] Fix standard -> LLVM conversion to fail for unsupported memref element type..

Would a test case be applicable here? I assume this leads to a pattern match failure and the op staying as is in a simple case.

Nov 12 2020, 12:44 PM · Restricted Project

Nov 11 2020

bondhugula added a comment to D90899: [mlir] Add DecomposeCallGraphTypes pass..

Is "call graph types" the best term here? This doesn't have to do with call graphs but with function signatures.(?) The older naming along the lines of FunctionSignatureConversion or DecomposeFuncSignature or UnpackFuncSignature looks clearer than anything with "CallGraph".

Nov 11 2020, 8:41 PM · Restricted Project

Nov 10 2020

bondhugula requested changes to D90589: [mlir] Use assemblyFormat in AllocLikeOp..
Nov 10 2020, 10:05 AM · Restricted Project

Nov 9 2020

bondhugula added inline comments to D90731: Add ElementwiseMappable trait and apply it to std elementwise ops..
Nov 9 2020, 10:02 PM · Restricted Project
bondhugula requested changes to D90991: [mlir][Linalg] Fuse sequence of Linalg operation (on buffers).
Nov 9 2020, 9:40 PM · Restricted Project
bondhugula requested changes to D90899: [mlir] Add DecomposeCallGraphTypes pass..

This appears nice to me. Some superficial comments to start with.

Nov 9 2020, 10:29 AM · Restricted Project
bondhugula added inline comments to D90798: [mlir][Affine] Refactor affine fusion code in pass to utilities.
Nov 9 2020, 10:13 AM · Restricted Project
bondhugula accepted D90798: [mlir][Affine] Refactor affine fusion code in pass to utilities.

This looks like a really great cleanup! Thanks! A list of minor comments.

Nov 9 2020, 10:04 AM · Restricted Project
bondhugula added inline comments to D90589: [mlir] Use assemblyFormat in AllocLikeOp..
Nov 9 2020, 5:58 AM · Restricted Project
bondhugula added inline comments to D90803: [MLIR] Support `global_memref` and `get_global_memref` in standard -> LLVM conversion..
Nov 9 2020, 5:54 AM · Restricted Project

Nov 7 2020

bondhugula accepted D88719: [mlir] Normalize the results of normalizable operations.

Looks good. Please address the remaining minor comments.

Nov 7 2020, 9:46 PM · Restricted Project

Nov 2 2020

bondhugula requested changes to D89990: [mlir] Remove notion of "decompose values" from BufferizeTypeConverter.

But this would break the mhlo to lmhlo conversion where the canonical form of an incoming TF model lowered to mhlo is to return a tuple of tensors, and that tuple gets unpacked into a list of memref output arguments on the function. If this functionality can be improved with a better separation of concerns, it may be good to do that. But we shouldn't delete this until then! The way it is now: the functionality is clearly defined with its own proper test cases. And it has an out of tree user in the mlir-hlo repo (-hlo-legalize-to-lhlo pass) - please don't remove this! :-)

Nov 2 2020, 9:25 PM · Restricted Project
bondhugula requested changes to D90589: [mlir] Use assemblyFormat in AllocLikeOp..
Nov 2 2020, 2:31 AM · Restricted Project

Oct 29 2020

bondhugula requested changes to D90411: TOSA MLIR Dialect.
Oct 29 2020, 11:58 AM · Restricted Project

Oct 23 2020

bondhugula resigned from D86071: [MLIR][OpenMP] Add omp.wsloop operation.
Oct 23 2020, 11:18 AM · Restricted Project, Restricted Project

Oct 21 2020

bondhugula added a reviewer for D88719: [mlir] Normalize the results of normalizable operations: avarmapml.
Oct 21 2020, 4:20 PM · Restricted Project
bondhugula requested changes to D88719: [mlir] Normalize the results of normalizable operations.
Oct 21 2020, 4:20 PM · Restricted Project

Oct 19 2020

bondhugula requested changes to D89720: [mlir] Add MemRefReinterpretCastOp definition to Standard..
Oct 19 2020, 8:13 PM · Restricted Project
bondhugula added a comment to D89720: [mlir] Add MemRefReinterpretCastOp definition to Standard..

Do you want to add a to in the syntax?

Oct 19 2020, 8:06 PM · Restricted Project

Oct 16 2020

bondhugula added inline comments to D89604: [mlir] Add canonicalization for cond_br that feed into a cond_br on the same condition.
Oct 16 2020, 9:55 PM · Restricted Project

Oct 13 2020

bondhugula requested changes to D87756: [mlir] Refactored BufferPlacement transformation into BufferDeallocation and BufferHoisting..

Mostly minor comments on readability.

Oct 13 2020, 12:50 AM · Restricted Project

Oct 10 2020

bondhugula committed rG9909ef292dad: [mlir][scf] Fix a bug in scf::ForOp loop unroll with an epilogue (authored by tatianashp).
[mlir][scf] Fix a bug in scf::ForOp loop unroll with an epilogue
Oct 10 2020, 1:49 AM
bondhugula closed D87583: [mlir][scf] Fix a bug in scf::ForOp loop unroll with an epilogue.
Oct 10 2020, 1:49 AM · Restricted Project
bondhugula accepted D87583: [mlir][scf] Fix a bug in scf::ForOp loop unroll with an epilogue.
Oct 10 2020, 1:44 AM · Restricted Project

Oct 9 2020

bondhugula added inline comments to D89153: [mlir] Fix TypeID for shared libraries built with -fvisibility=hidden..
Oct 9 2020, 12:08 PM · Restricted Project
bondhugula added inline comments to D89153: [mlir] Fix TypeID for shared libraries built with -fvisibility=hidden..
Oct 9 2020, 12:06 PM · Restricted Project
bondhugula added inline comments to D89112: [mlir][gpu] Fix bug in kernel outlining.
Oct 9 2020, 5:06 AM · Restricted Project
bondhugula added inline comments to D88341: [mlir][vector] First step of vector distribution transformation.
Oct 9 2020, 12:01 AM · Restricted Project

Oct 8 2020

bondhugula added a comment to D88341: [mlir][vector] First step of vector distribution transformation.
->

func @distribute_vector_add(%A: vector<32xf32>, %B: vector<32xf32>) -> vector<32xf32> {

%C = simt.execute %A to %Aelt : vector<1xf32> affine_map<()[s0] -> (s0)> 
                                 %B to %Belt : vector<1xf32> affine_map<()[s0] -> (s0)> {
  %0 = addf %A, %B : vector<1xf32>
  yield %0
}
Oct 8 2020, 11:53 PM · Restricted Project

Oct 6 2020

bondhugula added inline comments to D88941: [mlir][openacc] Introduce acc.enter_data operation.
Oct 6 2020, 9:56 PM · Restricted Project

Oct 4 2020

bondhugula accepted D87583: [mlir][scf] Fix a bug in scf::ForOp loop unroll with an epilogue.
Oct 4 2020, 11:26 PM · Restricted Project

Oct 3 2020

bondhugula resigned from D76821: Mass clang-format for MLIR codebase.
Oct 3 2020, 1:42 PM · Restricted Project, Restricted Project
bondhugula added inline comments to D88657: [mlir] Add subtensor_insert operation.
Oct 3 2020, 1:41 PM · Restricted Project
bondhugula resigned from D88651: Promote transpose from linalg to standard dialect.
Oct 3 2020, 1:33 PM · Restricted Project
bondhugula added inline comments to D88719: [mlir] Normalize the results of normalizable operations.
Oct 3 2020, 10:34 AM · Restricted Project
bondhugula added a comment to D88719: [mlir] Normalize the results of normalizable operations.

Thanks for adding this support. Some minor comments. I'll be able to review post the refactoring.

Oct 3 2020, 10:31 AM · Restricted Project
bondhugula requested changes to D88719: [mlir] Normalize the results of normalizable operations.
Oct 3 2020, 10:28 AM · Restricted Project

Oct 1 2020

bondhugula accepted D88532: [MLIR] Updates around MemRef Normalization.
Oct 1 2020, 9:00 PM · Restricted Project
bondhugula added a comment to D88532: [MLIR] Updates around MemRef Normalization.

A few more minor comments.

Oct 1 2020, 12:15 PM · Restricted Project
bondhugula added inline comments to D88651: Promote transpose from linalg to standard dialect.
Oct 1 2020, 8:05 AM · Restricted Project
bondhugula requested changes to D88651: Promote transpose from linalg to standard dialect.
Oct 1 2020, 8:03 AM · Restricted Project
bondhugula added a comment to D88532: [MLIR] Updates around MemRef Normalization.

Thanks for submitting this - esp. the detailed pass description. We should definitely improve the documentation since you didn't find it clear enough.

Oct 1 2020, 1:29 AM · Restricted Project
bondhugula requested changes to D88532: [MLIR] Updates around MemRef Normalization.
Oct 1 2020, 1:27 AM · Restricted Project
bondhugula added reviewers for D88532: [MLIR] Updates around MemRef Normalization: AlexEichenberger, avarmapml.
Oct 1 2020, 1:24 AM · Restricted Project

Sep 29 2020

bondhugula accepted D88168: [MLIR][Standard] Add `atan2` to standard dialect.
Sep 29 2020, 8:27 AM · Restricted Project
bondhugula requested changes to D88168: [MLIR][Standard] Add `atan2` to standard dialect.
Sep 29 2020, 3:21 AM · Restricted Project
bondhugula added inline comments to D88168: [MLIR][Standard] Add `atan2` to standard dialect.
Sep 29 2020, 3:21 AM · Restricted Project

Sep 25 2020

bondhugula requested changes to D88168: [MLIR][Standard] Add `atan2` to standard dialect.
Sep 25 2020, 11:44 AM · Restricted Project
bondhugula committed rGc1f856803142: [MLIR] Fix for updating function signature in normalizing memrefs (authored by imaihal).
[MLIR] Fix for updating function signature in normalizing memrefs
Sep 25 2020, 10:28 AM
bondhugula closed D87746: [mlir] Fix of updating function signature in normalizing memrefs.
Sep 25 2020, 10:28 AM · Restricted Project, Restricted Project
bondhugula accepted D87746: [mlir] Fix of updating function signature in normalizing memrefs.
Sep 25 2020, 10:18 AM · Restricted Project, Restricted Project

Sep 24 2020

bondhugula committed rGff00b5839252: [MLIR] Normalize memrefs in LoadOp and StoreOp of Standard Ops (authored by imaihal).
[MLIR] Normalize memrefs in LoadOp and StoreOp of Standard Ops
Sep 24 2020, 6:31 AM