Aart J.C. Bik received his PhD from the Leiden University in 1996. From 1998 to 2007, he worked at Intel, where he was the lead compiler architect of automatic vectorization in the Intel C++/Fortran compilers. In 2007, he moved to Google, where he has worked on compilers for ART and the Dart VM. He is now a member of the MLIR codegen team, focusing on compiler support for sparse tensor computations.
User Details
User Details
- User Since
- Jan 10 2020, 10:13 AM (124 w, 22 h)
Yesterday
Yesterday
aartbik added inline comments to D126570: [mlir][bufferization] Add optional `copy` operand to AllocTensorOp.
aartbik added a comment to D126180: [mlir][sparse][NFC] Switch InitOp to bufferization::AllocTensorOp.
I really like the direction this is going in, so we have one centralized mechanism to reason on allocation (and deallocation) of the buffers for tensors!
In the long run, it would be nice to have bufferization.de_alloc_tensor() as well, and have one mechanism that inserts those, independent of dense/sparse tensors.
aartbik added a comment to D126333: [mlir][sparse] Add optional start_value to sparse_tensor.init.
Note that there is this parallel effort to unify our tensor initialization ops (see https://reviews.llvm.org/D126180).
So let's brainstorm offline to come up with another solution.
[OpenMP][mlir] fix broken build
aartbik requested review of D126556: [OpenMP][mlir] fix broken build.
Thu, May 26
Thu, May 26
aartbik added inline comments to D126333: [mlir][sparse] Add optional start_value to sparse_tensor.init.
aartbik updated the diff for D126422: [mlir][python][ctypes] fix ctype python binding complication for complex.
rebased with main
aartbik added a comment to D126422: [mlir][python][ctypes] fix ctype python binding complication for complex.
aartbik updated the diff for D126422: [mlir][python][ctypes] fix ctype python binding complication for complex.
added tests for c128 and c64, ranked and unranked, memrefs
Wed, May 25
Wed, May 25
aartbik updated the summary of D126422: [mlir][python][ctypes] fix ctype python binding complication for complex.
aartbik updated the summary of D126422: [mlir][python][ctypes] fix ctype python binding complication for complex.
aartbik requested review of D126422: [mlir][python][ctypes] fix ctype python binding complication for complex.
Onward to feature completeness!
aartbik added a comment to D126409: Lower complex.expm1 to standard dialect..
Very neat. The general idea LGTM but I leave to it our math experts to approve.
Tue, May 24
Tue, May 24
[mlir][sparse] complex lowering
aartbik updated the diff for D126335: [mlir][sparse] complex lowering.
added CMake
aartbik requested review of D126335: [mlir][sparse] complex lowering.
aartbik committed rG5799f843a220: [mlir][sparse] add new complex ops to reduction recognition (authored by aartbik).
[mlir][sparse] add new complex ops to reduction recognition
aartbik requested review of D126318: [mlir][sparse] add new complex ops to reduction recognition.
aartbik added inline comments to D126105: [mlir][sparse] Improving ExecutionEngine/SparseTensorUtils.h.
Fri, May 20
Fri, May 20
Excellent work, Bixia!
aartbik added a comment to D126027: [mlir][sparse] Support more complex operations..
Nice progress. One comment.
aartbik committed rG28b6d412afc5: [mlir][sparse] add support for complex zero/one building (authored by aartbik).
[mlir][sparse] add support for complex zero/one building
Thu, May 19
Thu, May 19
aartbik added reviewers for D126039: [mlir][sparse] add support for complex zero/one building: bixia, wrengr.
aartbik requested review of D126039: [mlir][sparse] add support for complex zero/one building.
Oh, nice, I wonder if that makes reading the very large tensors slightly faster again!
aartbik committed rG3b13f8805c02: [mlir][sparse] fix unsigned comparison bug in assert (authored by aartbik).
[mlir][sparse] fix unsigned comparison bug in assert
aartbik requested review of D126007: [mlir][sparse] fix unsigned comparison bug in assert.
Tue, May 17
Tue, May 17
aartbik accepted D125834: [mlir][sparse] Add support for complex.im and complex.re to the sparse compiler..
Nice! Very fast. A few nits, but good to go.
Mon, May 16
Mon, May 16
Good to go after the required rebasing and merging.
I like this a lot! You will have to rebase for my complex change (sorry), but good to submit without review after that.
aartbik committed rG736c1b66ef33: [mlir][sparse] introduce complex type to sparse tensor support (authored by aartbik).
[mlir][sparse] introduce complex type to sparse tensor support
aartbik updated the diff for D125596: [mlir][sparse] introduce complex type to sparse tensor support.
correct include for complex type
aartbik updated the diff for D125596: [mlir][sparse] introduce complex type to sparse tensor support.
addressed comments
aartbik added inline comments to D125596: [mlir][sparse] introduce complex type to sparse tensor support.
aartbik updated the diff for D125596: [mlir][sparse] introduce complex type to sparse tensor support.
CMake fix
Fri, May 13
Fri, May 13
aartbik requested review of D125596: [mlir][sparse] introduce complex type to sparse tensor support.
Jumping a bit ahead in the stack, but this delta is good to go once base goes in as well.
Thu, May 12
Thu, May 12
aartbik committed rG6f3c7dfb7746: [mlir][sparse] add sparse sign integration test (authored by aartbik).
[mlir][sparse] add sparse sign integration test
aartbik updated the diff for D125494: [mlir][sparse] add sparse sign integration test.
discussed offline: only one padding element
aartbik updated the diff for D125494: [mlir][sparse] add sparse sign integration test.
TAB -> spaces
aartbik retitled D123876: [mlir][sparse] Expose SparseTensor passes as enums instead of opaque numbers for vectorization and parallelization options. from [mlir][sparse] Expose SpareTensor passes as enums instead of opaque numbers for vectorization and parallelization options. to [mlir][sparse] Expose SparseTensor passes as enums instead of opaque numbers for vectorization and parallelization options..
aartbik added a comment to D125501: [mlir] Remove the `type` keyword from type alias definitions.
LGTM for sparse
aartbik added reviewers for D125494: [mlir][sparse] add sparse sign integration test: wrengr, jim22k.
aartbik requested review of D125494: [mlir][sparse] add sparse sign integration test.
aartbik accepted D122060: [mlir][sparse] Factoring out an enumerator over elements of SparseTensorStorage.
Thanks for your patience during the review, Wren. It has been a long road, but nice to see this new abstraction!
Changes look good, and it is of course good to require the minimum required type. But is there a major advantage of doing it this way? (follow up or otherwise)?
Wed, May 11
Wed, May 11
aartbik added a comment to rG298d2fa1c5f1: Apply clang-tidy fixes for readability-identifier-naming in SparseTensorUtils..
Thanks for the fix. I always fall back to my 90ties C style ;-)
The reported failure seems unrelated but please double-check too.
Thanks for fixing this so quickly!
Fri, May 6
Fri, May 6
aartbik added a comment to rG6a9c1029f845: Fix build with shared libs: add missing CMake dep to MLIR sparse pipeline.
Thank you kindly, Mehdi, as always, for watching our BUILD so closely and proactively fixing it!
aartbik committed rG5b122a7310e8: [mlir][sparse] integration test for zero preserving math op (authored by aartbik).
[mlir][sparse] integration test for zero preserving math op
aartbik requested review of D125104: [mlir][sparse] integration test for zero preserving math op.
Thu, May 5
Thu, May 5
aartbik committed rG952fa3018e73: [mlir][sparse] add more zero-preserving unary ops to sparse compiler (authored by aartbik).
[mlir][sparse] add more zero-preserving unary ops to sparse compiler
aartbik added a comment to D122060: [mlir][sparse] Factoring out an enumerator over elements of SparseTensorStorage.
Do you have some experimental validation to report before we proceed with this?
aartbik requested review of D125043: [mlir][sparse] add more zero-preserving unary ops to sparse compiler.
Wed, May 4
Wed, May 4
Nice! Thank you for doing this so quickly!
Tue, May 3
Tue, May 3
aartbik committed rG2617f2f70826: [mlir][sparse] fix build issue with unused local under opt builds (authored by aartbik).
[mlir][sparse] fix build issue with unused local under opt builds
aartbik added reviewers for D124883: [mlir][sparse] fix build issue with unused local under opt builds: jim22k, bixia, wrengr.
aartbik requested review of D124883: [mlir][sparse] fix build issue with unused local under opt builds.
aartbik committed rG1abcdc677cfd: [mlir][sparse] add missing types to from/to-MLIR conversion routines (authored by aartbik).
[mlir][sparse] add missing types to from/to-MLIR conversion routines
aartbik requested review of D124875: [mlir][sparse] add missing types to from/to-MLIR conversion routines.
Mon, May 2
Mon, May 2
A few last comments, but I am giving you the LGTM, since I think this is ready to go in, so we can work on refining the few remaining issues.
Thanks for your patience during the review, and thanks for this contribution!
Thu, Apr 28
Thu, Apr 28
aartbik added inline comments to D123057: [mlir][sparse] Lowering for unary and binary.
Apr 27 2022
Apr 27 2022
aartbik added a comment to D123057: [mlir][sparse] Lowering for unary and binary.
A few last comments. Sorry for being nitpicky on this, rest assured I really like the work, I am just a bit particular on certain things.
Also, can you please make a pass over all my comments and mark them "Done" (or comment on them otherwise).
That makes the re-review a bit easier, since a lot of my past comments still show up as unresolv.ed
aartbik added a comment to D123876: [mlir][sparse] Expose SparseTensor passes as enums instead of opaque numbers for vectorization and parallelization options..
[mlir][sparse] optimize COO index handling
aartbik updated the diff for D124502: [mlir][sparse] optimize COO index handling.
added note on change to struct