Ecosyste.ms: OpenCollective

An open API service for software projects hosted on Open Collective.

github.com/llvm/torch-mlir

The Torch-MLIR project aims to provide first class support from the PyTorch ecosystem to the MLIR ecosystem.
https://github.com/llvm/torch-mlir

Bump llvm-project to c8c07b76b2cf2ada8e7ec132f7f57b97d76743cf.

* Several NFC changes to signatures/includes.

a3f4db9fe8d9574c9948f1662bf7d681cc5ecdcc authored about 4 years ago by Stella Laurenzo <[email protected]>
Fix TCP_MulOp tablegen definition

30adf9e6b0c1e94db38050a9e143f20a5a461d17 authored about 4 years ago by Marius Brehler <[email protected]>
Rewrite ATen ODS code generator to be based on new op registry and new signature recognition system.

* Deletes prior code generator from previous attempt (moved some of it into this one).
* Renames...

c08935a4182d2333e9dd71d35bad635a15edf916 authored about 4 years ago by Stella Laurenzo <[email protected]>
[RefBackend] Support element-wise multiply op

Register the following for the multiply op:
- tcf.mul
- tcp.mul
- TCP->TCP lowering
- Shape tran...

94ea6f7c929b37ba9aafd49fe6dca4a2f9374ba1 authored about 4 years ago by Aaron J Arthurs <[email protected]>
Expose signature metadata to ops and implement ATenRecognizeKernelsPass pass.

* Two op interfaces, one for querying instance metadata and one for getting static data needed t...

510f226df27448c6f49bff49809051d0bf37e422 authored about 4 years ago by Stella Laurenzo <[email protected]>
Add missing dependency on NPCOMPCAPI from NPCOMPPythonCommon

Fix linker error:

lib/Python/libNPCOMPPythonCommon.a(MlirInit.cpp.o): in function `mlir::npcomp...

f3c75d957b5146b5a53dd2dd53782a0d5b8d6277 authored about 4 years ago by Mehdi Amini <[email protected]>
NFC: Transition ATen passes to tablegen registration.

91fc83d2e78fcbc05ac24876790b82ad68189e68 authored about 4 years ago by Stella Laurenzo <[email protected]>
NFC: Re-organize ATen directory structure and fix warnings.

* Still some more work to do on the Transforms tree to bring it in line with the others (will do...

9618c2dbf76d6a4a06cdd17ba80e007154288ecc authored about 4 years ago by Stella Laurenzo <[email protected]>
NFC: Use new print with large_elements_limit in tests.

* For tests with large constants, decreases issues with lit pipelines.
* Bumps llvm-project to p...

d09300886a219ad0e46f8f9266759cd50c9f85fb authored about 4 years ago by Stella Laurenzo <[email protected]>
[RefBackend] Use upstream std bufferization.

It now subsumes the one we had.

14470f9ff6f7de337debe367887a8e70a7dceab7 authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Use new upstream SCF type conversions.

b6ae53b312173399eccc3f1584575658f5903c4f authored about 4 years ago by Sean Silva <[email protected]>
Bump llvm-project to 57b338c08a4942bda6e58c77870c657c53b6fb5b

Date: Mon Oct 19 15:59:03 2020 -0700

4f2aa12d1a680280b7ae7d7d7b720c877ba1e45b authored about 4 years ago by Sean Silva <[email protected]>
NFC: Format a file that had not been.

fe5ceed18d0926f32676bd3516df11390a8a6f4b authored about 4 years ago by Stella Laurenzo <[email protected]>
Work around various PyTorch issues in support of convolution.

* Enables the conv2d fwd test and ResA (which are both small).
* Deletes resnet18 and vgg, which...

58adb6bd8e12b0eaecac07c6e3e72e8968fa657d authored about 4 years ago by Stella Laurenzo <[email protected]>
Add remaining pieces to capture full example models.

* Adds Basicpy List, Tuple, Dict types and plumbs through C API.
* Started debugging the issues ...

029815152e6a072e3a7f8d817accaae18d6877e1 authored about 4 years ago by Stella Laurenzo <[email protected]>
Only build mnist-playground if Torch was found.

81119aa0a10d1ee759ac0641e406b89a793a7a48 authored about 4 years ago by Sean Silva <[email protected]>
Replace deprecated add-path in GitHub Action

The add-path command is deprecated and should be replaced as stated in
https://github.blog/chang...

54c0a3a1e9a5451b54c8e651c8d7e3142521a7d2 authored about 4 years ago by Marius Brehler <[email protected]>
More progress on PyTorch acap device capture.

* Now gets far enough to capture batch_norm.
* Has some issues still with in-place ops.
* Can ma...

9e52f6235be6c636fe3a29a743c526dc9641941c authored about 4 years ago by Stella Laurenzo <[email protected]>
[build_tools] Fix install_mlir.sh to look for Python3

Otherwise, on my machine MLIR somehow gets configured with Python 2,
which is not supported.

Th...

87b6bc305a528c713ced372e3bba39b84e3ed80d authored about 4 years ago by Sean Silva <[email protected]>
run_lit.sh: build all the things

npcomp-capi-ir-test was missing, causing the tests to fail if it wasn't
built. Just build everyt...

ebb3c9035550518697fb809ad7dc1ee971325c29 authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Use more idiomatic bufferize pattern for TCP.

The time has come for BypassShapes/LowerShapedResultsToMemref to go away :(
For the reference ba...

06a8ba69005d59dc7ec02598424375db7835dc9d authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Use upstream BufferizeTypeConverter

Now that it has grown source/target materialization capabilities
(spelled with ops tensor_load/t...

b6bdc8cc4fcb555510901148afbc222735026662 authored about 4 years ago by Sean Silva <[email protected]>
Fix up bash_helpers.sh for new build directory layout.

Also, remove the PYTHONPATH stuff for now, I think it is out of date and
I haven't been able to ...

a7a1d1be2a98afd8b5761e9b18da9a6b1bc53979 authored about 4 years ago by Sean Silva <[email protected]>
Update llvm-project to ee491ac91e123b90eeec3cce7e494936ea8cb85d

Date: Wed Oct 14 11:26:22 2020 -0700

1d118b0845b1af70d6e8bff95ead4b661eb62ddd authored about 4 years ago by Sean Silva <[email protected]>
Port prior acap export tests to new dispatcher based versions.

* Sadly, non-trivial ones fail.
* Bugs filed and marked XFAIL.

abb6fe8aa29939b179652e7b0d3594e298b90964 authored about 4 years ago by Stella Laurenzo <[email protected]>
Create public API for torch_mlir python code.

* Adds a trampoline/loader 'torch_mlir' module.
* Plumbs through the MLIR python Context and Mod...

30cfc6499f1fa937e558112715e7b51f39683150 authored about 4 years ago by Stella Laurenzo <[email protected]>
Enable MLIR python bindings in CI.

86df4cabeb3468ea2d30297aedf77c5fc30b68c7 authored about 4 years ago by Stella Laurenzo <[email protected]>
Update test configuration to import mlir from LLVM install location.

* Also adds two lit tests to verify that all of our extensions load without fireworks, which is ...

5c5b8db70ff3c2ce6b231a95b031e7147b289e59 authored about 4 years ago by Stella Laurenzo <[email protected]>
Bump llvm-project to 820e65f9e2369d2990fde4b3e7cfceb64f0df9c8

Date: Mon Oct 12 11:26:50 2020 -0700

f2d5c26c977a18f66df7ba014594f36f45a55aa7 authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Split out TCF->TCP conversion.

Now the reference backend is cleanly accepts "TCP"+scalar ops.

We introduce tcf-refback-lowerin...

93fc21dad04cff3a571a587bde736a85e0b4e51a authored about 4 years ago by Sean Silva <[email protected]>
Add another config var to the CI for shared lib builds.

Missed this one in the previous change.

83342e871562ece41e38044a175fa171773a8fea authored about 4 years ago by Stella Laurenzo <[email protected]>
Start reworking towards a shared library build.

* Need to have a dag of shared library deps in order to interop across python extensions (as pre...

af4edb63aef9294d08c07cb3843c0e308c12d97b authored about 4 years ago by Stella Laurenzo <[email protected]>
Enable libMLIR.so in CI.

d9dc16a9be5df3b1bfb89834cb8d85a45036fc52 authored about 4 years ago by Stella Laurenzo <[email protected]>
Fix up the docker script / instructions after attempting to use it.

d6b05c507ad51b11bc676357e1e28fa8c8a76f03 authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Rename RefBackend dialect to Refback

I now realize that VerboseCamelCase is not the best choice for dialect
directory/file names and ...

7edb5f364179b57c75941ce3e7340cf1ce8c1158 authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Rename Npcomprt dialect to Refbackrt.

bf99a82832f84ccc1d526e8d50423ae85b699524 authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Put JITModule in refback namsepace.

631c8070dfefc666682a5d2542c942413af4cc1c authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Move runtime related code under npcomp/RefBackend/

Other than the dialect definitions (which will live in standard Dialect/
subdirectory), the goal...

83ad70ef548a188277d7c77f6874ee38d5bf7df5 authored about 4 years ago by Sean Silva <[email protected]>
Add scripts/documentation for VSCode setup with a docker dev image.

* Forks a subset of my shell functions into docker_shell_funcs.sh, specifically needed to create...

51d51241b4860b6c3a5a9fe9fae16edeebbcc0cc authored about 4 years ago by Stella Laurenzo <[email protected]>
[RefBackend] Rename test/E2E.

I missed this in the previous commits.

ddc2e9de5d83f6c25260ac3981d6786b74f64574 authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Split out RefBackend (refback) dialect from TCP.

This is the first in a patch series that is refactoring the
constellation of things variously ca...

5017430dc78de1d074d1b1c2cc90685bfecbbe9d authored about 4 years ago by Sean Silva <[email protected]>
[RefBackend] Rename "E2E" to RefBackend.

21255d5f8e7d6db2a20edf65c07fdbf253986456 authored about 4 years ago by Sean Silva <[email protected]>
Rename a couple CMake targets.

NPCOMPFoo to NPCOMPFooDialect for consistency with others.

03846ed8e734590ca1b03cd76c8dbdee2eaf1523 authored about 4 years ago by Sean Silva <[email protected]>
Set PyTorch captured function return type.

* Resolves various TODOs that required an LLVM change/bump.
* Bumps LLVM to 4aa217160e5f06a96c6e...

3ccc2214a7adadb1307683b872f8b8dbdc677f40 authored about 4 years ago by Stella Laurenzo <[email protected]>
Bump llvm to ed46e84c7aaffd847656ac559acb06089096ec33.

* Minor change of MLIRStandardOps -> MLIRStandard

58b6033537a518758d6d3364da2d0a28df882feb authored about 4 years ago by Stella Laurenzo <[email protected]>
Implement torch.kernel_call capture.

* Had to stop short of modifying the function return signature because of a missing C-API upstre...

ad3ddb9edbb40d7c8866c6e2eb74f1edfedf76d6 authored about 4 years ago by Stella Laurenzo <[email protected]>
Add hopefully short-lived mnist-playground utility.

This unblocks backend progress while the PyTorch frontend work is coming
online. Hopefully we ca...

dd1fa2607faf50a80714eaada5104f71a353c0f7 authored about 4 years ago by Sean Silva <[email protected]>
[RefE2E] Initialize the linalg matmul accumulator buffer.

I was seeing some miscompiles due to the uninitialized data read here
before. Interestingly, thi...

8022dfaf1a39dc2e4f8e57419a0b737bf191c7e2 authored about 4 years ago by Sean Silva <[email protected]>
Add capture function arguments.

* Adds at::Tensor -> MlirValue tracking.
* Adds conversions for tensor and scalar types to MLIR ...

e5433e314f5113aa5485ef4ed6040739b2968edd authored about 4 years ago by Stella Laurenzo <[email protected]>
Add a torch.kernel_call op and associated predicates.

3d74337be0443440402e2419ce07c9c0ee77baf5 authored about 4 years ago by Stella Laurenzo <[email protected]>
Add public API for constructing a module/function to capture PyTorch ops.

* Uses the MLIR-C API since that will save us a lot of grief down the road (i.e. will give PyTor...

ba03ecc652adb94489c1fd258b8df7384119f210 authored about 4 years ago by Stella Laurenzo <[email protected]>
Bump LLVM to e72d792c147ee506e337401e20c0f23042cc43fe.

* Does not bump mhlo as an upstream integrate on that project has not taken place and there is n...

9722a6ce781907c2313eb0e7fc3224c386fca4d9 authored about 4 years ago by Stella Laurenzo <[email protected]>
Add boilerplate for Torch dialect.

2c9ca79c8945dc8426fcfede34d4c980e192eb24 authored about 4 years ago by Stella Laurenzo <[email protected]>
Run format_sources.sh.

fb895173f291afaaa9a4cb5a80daa6d0983dc169 authored about 4 years ago by Stella Laurenzo <[email protected]>
Add boilerplate to do device capture (pytorch 1.6).

* Uses the new dispatcher API.
* Just prints to the console for the moment when an op is capture...

b5f010284f4fa4bfeae94695f7878dc2912de7ac authored about 4 years ago by Stella Laurenzo <[email protected]>
[RefE2E] Use upstream shape constraint conversion pass.

Now that we upstreamed our pass, we can remove it.
The final pass that landed upstream doesn't d...

16c26ef57e8272254963e660106d9e5fc566c5a1 authored about 4 years ago by Sean Silva <[email protected]>
Bump llvm-project to 9ed1e5873c19eb817fb9e36d0262c7effee5d35e

Date: Fri Sep 18 13:55:52 2020 -0700

- Update to linalg syntax
- New generated builders are b...

6ea37cfed69def885028ab715ab4b121e7e1cdd5 authored about 4 years ago by Sean Silva <[email protected]>
[RefE2E] Add support for unary ops exp and tanh

This is fairly mechanical.

f9b37c55b775d6fec70152db9f08ad7bbe353eff authored about 4 years ago by Sean Silva <[email protected]>
[NFC] Remove stray .dump() that snuck in.

6b69beae6aceafd534b15ad77f8c704cc7c065b2 authored about 4 years ago by Sean Silva <[email protected]>
Move tests around so we can have dedicated tests for the c10 dispatcher.

* Adds a trivial missing test for _torch_mlir.c10.get_registered_ops()
* Disables the regression...

0cb28f0b06fd512f6ff32e81781630a969ec5f1a authored about 4 years ago by Stella Laurenzo <[email protected]>
Add compatibility notes regarding unpacking quantized weights. (#56)

Co-authored-by: Bryce Arden <[email protected]>

6e6efb2854dca4a41b55c2de9e3390ddad30dbb4 authored about 4 years ago by Stella Laurenzo <[email protected]>
Add docker image/instructions for building against pytorch 1.6.

47c3a9f461ae775259073f37c57d397bf96fc25d authored about 4 years ago by Stella Laurenzo <[email protected]>
Add initial python bindings for c10 dispatcher internals. (#55)

* Exposes the op registry via a get_registered_ops method.
* Moves the aten dialect generation ...

0d9188596553e1129fddc94345fc821c8edecfc8 authored about 4 years ago by Stella Laurenzo <[email protected]>
[RefE2E] Add support for "max".

This cleans up the lowering pipeline to easily allow extending to
multiple binary ops. It looks ...

c69e9fabc5c3b148722335610fe6844f9bdbe137 authored over 4 years ago by Sean Silva <[email protected]>
Inject missing dialects in E2E passes

681c4e1d4ac879921a852e6243288fd159de222a authored over 4 years ago by Marius Brehler <[email protected]>
[RefE2E] Add interesting control flow example.

This also required adding a lowering for ForOp in our tensor->memref
conversion.

7b7f35744bb9ff9a492afb237fa5bd60160833b4 authored over 4 years ago by Sean Silva <[email protected]>
Add more ops from the original integration.

* Still need to add a systematic mechanism for discovering gradient ops.
* Work needed on the va...

bc7c8523796f20a92bce81b4400f538b97e1d9d6 authored over 4 years ago by Stella Laurenzo <[email protected]>
[RefE2E] Refactor how tcf.add is lowered.

It was previously going through this awkward route that prematurely
created linalg.generic ops, ...

dc8afc927136fcccc4a89e8e0de092b57ebcaf2f authored over 4 years ago by Sean Silva <[email protected]>
[RefE2E] Add assemblyFormat for TCF and TCP ops and tidy up.

276f5b80ea2fea4251d7f25b4725243fba58715b authored over 4 years ago by Sean Silva <[email protected]>
Fixes for run_lit.sh

- new build directory layout
- build NPCOMPNativePyExt, now that lit tests use it

7486befffd9d1a5a6b03b7f5db7b49a8e68bbce4 authored over 4 years ago by Sean Silva <[email protected]>
[RefE2E] Add support for matmul.

I'm pretty happy with how this turned out. It looks pretty much like it
should -- one change at ...

d8675f8ad27ba410d800cdca080195909c83e5b9 authored over 4 years ago by Sean Silva <[email protected]>
[RefE2E] Fix nul-termination bug.

I was seeing some of the error messages come out with some garbage at
the end. This fixes it.

62738d3641ce3c7b10775b170b1287b84d6838c1 authored over 4 years ago by Sean Silva <[email protected]>
Bump llvm-project to 7c44651360dd94e17011fd1cd7ec3c755e0363b4

Date: Thu Sep 17 18:16:41 2020 -0700

2284f6b4f105a3d84ff7819973642c593ba95b49 authored over 4 years ago by Sean Silva <[email protected]>
Update README to reference published docker tag.

361abebb517087a388feb9bb66705c15e981203a authored over 4 years ago by Stella Laurenzo <[email protected]>
Explicitly load aten and std dialects when constructing a context. (#47)

* This gets the pytorch frontend broadly working and what is left appears to be legitimate failu...

8ac29594df3782dcc4fd14c5342e80831c90ea4b authored over 4 years ago by Stella Laurenzo <[email protected]>
Update docker, instructions and some fixes for the pytorch 1.3 build. (#45)

* Includes pybind11 directly (for some reason using the pytorch helper header for this depends o...

678989a321eaf7905dfa7788759946227c7c8bc0 authored over 4 years ago by Stella Laurenzo <[email protected]>
Totally rework RefE2E tensor to memref flow. (#42)

This now gets the overall "RefE2E" compilation stack to a point that I'm
fairly happy with. We ...

75f57b461e18597c9bd7b0a862d32cbcd0f9dfdf authored over 4 years ago by Sean Silva <[email protected]>
Add a new python script to auto-generate ATen op ODS definitions. (#43)

* Add a new python script to auto-generate ATen op ODS definitions.

* There is still some wor...

a74a98094b88d2e8b4d532ffe711f3fa94a56ccb authored over 4 years ago by Stella Laurenzo <[email protected]>
Bump LLVM to @7d1ed69 and fix namespace handling changed upstream.

* Bump LLVM to llvm/llvm-project@7d1ed69
* Bump MLIR-HLO to tensorflow/mlir-hlo@1880f87
* Adop...

d62f8227c2a205931158a59afd0c0c4e1e15760a authored over 4 years ago by Marius Brehler <[email protected]>
Run clang-format on files that do not comply.

dd9172fd75f0c12b8f6b6c45d12fed12c65c5773 authored over 4 years ago by Stella Laurenzo <[email protected]>
Bump llvm-project to 84a6da67e6b2a76b15ad1862f4cbb7625fe318df

That commit is a Thu Sep 10 22:04:58 2020 -0700

That change is required for a PR that I'm going...

0f9c6b4a35e7d1f393c50d3e808f6da42dbfdffb authored over 4 years ago by Sean Silva <[email protected]>
Register dialects in E2E passes

843448cde9aacbd27786f6b64b6dd7013f938656 authored over 4 years ago by Marius Brehler <[email protected]>
Remove unused include

a2fb68059fd3a62ee01e644471a7442fd4a09d02 authored over 4 years ago by Marius Brehler <[email protected]>
Register dialects in ATen lowering pass

124bc65a70dff7635e576daab0748a4ab5fa3beb authored over 4 years ago by Marius Brehler <[email protected]>
Register dialects in conversion passes

fb2d1a1559a9bbe374f618b60e05c6620d153ab1 authored over 4 years ago by Marius Brehler <[email protected]>
Integrate upstream LLVM at 8d9c13f37d2081c11186718ae8b5aef8b507d152.

* mlir-hlo: 062a3ac4a0671d15b5199ed2cd3a9ce02a5bf077

Fixes:

* numInputs() just returns an int ...

81dd571c23d489e7cc59369bda484821f6f13d3c authored over 4 years ago by Stella Laurenzo <[email protected]>
Bump submodule versions.

* llvm-project: b5924a8e27536d19dd5c4d302db29fb6163d5faa
* mhlo: 848ca244d20f045b7921da55a98a04d...

97d83f786a19281cd0fef5544c79e2ac420812c3 authored over 4 years ago by Stella Laurenzo <[email protected]>
Update build instructions to use the submodule for llvm.

* Previous instructions were referring to the option to use an external llvm-project checkout wi...

4c37aed8411a495effac1e924496719e13579cfa authored over 4 years ago by Stella Laurenzo <[email protected]>
Initial work on a torch op registry.

* This extracts metadata from python invocations (nearly) sufficient to generate ODS and a Torch...

d1ed6d260e77d332f631b33cb762110377676f15 authored over 4 years ago by Stella Laurenzo <[email protected]>
Format sources.

fc4f3743452fc02abf326350b18e360b4a6cde94 authored over 4 years ago by Stella Laurenzo <[email protected]>
Make code that depends on the legacy "type dispatch" mechanism optional. (#32)

* Make code that depends on the legacy "type dispatch" mechanism optional.

* This code is fai...

de38caa54713d50496554c7b8cdb3634eebc4d91 authored over 4 years ago by Stella Laurenzo <[email protected]>
Add pytorch interface to ATen Dialect (#30)

This patch adds a pytorch interface to npcomp. This interface is modeled
after pytorch_xla and...

31b3041e881dbc68f0b9c17bca83067f49ba5451 authored over 4 years ago by stephenneuendorffer <[email protected]>
NFC: Fix extra namespace declaration.

* Was causing build break on GCC9.

69cda404ef285bf13960cef6bbcae0aeab2c7eab authored over 4 years ago by Stella Laurenzo <[email protected]>
Create frontends/pytorch directory. (#31)

* Adds/updates readmes with some notes about code organization and direction.
* Meant to prepar...

77b235f6211a505a74150a57317470f4af7044ad authored over 4 years ago by Stella Laurenzo <[email protected]>
Add mlir-hlo as a submodule and add a script to find versions. (#20)

* I expect that mlir-hlo will be a non-optional dependency of the project, so adding as a sub-mo...

a2a36aa8f335517e6128ef47c518816fb37e6e8c authored over 4 years ago by Stella Laurenzo <[email protected]>
Add ATen Dialect (#16)

This patch adds a dialect intended to be used as a frontend dialect
to facilitate lowering from...

bb668e6e261a83da8b4b377b13405aef62c039a9 authored over 4 years ago by stephenneuendorffer <[email protected]>
Move precommit to 20.04 (#15)

14f614396df87bcfda393e238b2bee129b8e3be7 authored over 4 years ago by stephenneuendorffer <[email protected]>
Fix build again (#14)

The RuntimeShlib.so now lives in /lib.

5beaf4cc012fe87073378d607f127e08a27b48f3 authored over 4 years ago by stephenneuendorffer <[email protected]>
Fix precommit workflow (#13)

a5f3b16f92968f48ac4f549cf1ad73444ea650f2 authored over 4 years ago by stephenneuendorffer <[email protected]>
Fix build error (#12)

This debug rule only works with add_mlir_library.

111ba12e7f0666157505520b02f623adcab9917d authored over 4 years ago by stephenneuendorffer <[email protected]>