Commit Graph

459 Commits

Author SHA1 Message Date
Yan Da
c7ad928e60 More progress on WhileOp codegen 2022-04-05 15:55:48 +08:00
Yan Da
0f96da336a codegen for If 2022-04-04 12:58:37 +08:00
Yan Da
9df899b291 Some progress on visit_If 2022-04-03 22:34:46 +08:00
Yan Da
c71c50cd0c ForOp's SSA construction 2022-04-03 19:11:47 +08:00
Yan Da
61413b8a97 More python bindings 2022-04-01 22:22:39 +08:00
Yan Da
9dafa0e2e3 Update trtion dependencies 2022-04-01 20:16:07 +08:00
Yan Da
bde103fab0 Replace MlirType with mlir::Type 2022-04-01 18:46:46 +08:00
Yan Da
4ad432f1fc More on scf Ops 2022-03-31 21:42:48 +08:00
Yan Da
2041b67fbf Now vecadd works 2022-03-30 20:21:47 +08:00
Yan Da
e381dc72c5 Use mlir::Block to replace MlirBlock 2022-03-30 16:31:03 +08:00
Yan Da
e95d98a886 bindings for ModuleOp 2022-03-30 13:32:52 +08:00
Yan Da
38e67b4293 Add more Ops 2022-03-28 19:50:23 +08:00
Yan Da
0d139ec460 Introducing SCF 2022-03-26 17:02:32 +08:00
Yan Da
c53f3486e4 create shr 2022-03-26 16:41:49 +08:00
Yan Da
ba16116f96 Let python manage created objects 2022-03-26 16:31:01 +08:00
Yan Da
fed9925bbd Using stable LLVM release 2022-03-26 16:25:18 +08:00
Yan Da
a17fba86b1 Logic Op creation 2022-03-26 16:16:20 +08:00
Yan Da
d5612333c0 More fcmp ops 2022-03-25 14:12:20 +08:00
Yan Da
07881b4d41 Update includes 2022-03-24 13:46:35 +08:00
Yan Da
cf7fc8d642 Update includes 2022-03-24 13:33:54 +08:00
Yan Da
14a71dcb6f Replace MlirOperation with MlirValue 2022-03-23 13:31:14 +08:00
Yan Da
f2ab318614 New python binding 2022-03-22 21:53:22 +08:00
Yan Da
419bbe0f6e Reverts back to MLIR 14 & updates CMakeLists 2022-03-20 16:41:48 +08:00
Yan Da
a2c31ff434 Init commit 2022-03-17 20:40:55 +08:00
daadaada
539961072c [FRONTEND] Semantic analysis refactor (#473)
Moved dispatch.cc to semantic.py
Integer signedness now moved from C++ to python
Cleaner frontend type

Co-authored-by: Phil Tillet <phil@openai.com>
2022-03-16 21:25:30 -07:00
Yongjik Kim
0dd2ec2e3a [FRONTEND] Add an assert in case we get a CPU tensor. (#478) 2022-03-16 14:38:56 -07:00
Philippe Tillet
d4d8eaf6c0 [FRONTEND] improved caching mechanism (#474)
Co-authored-by: Greg Brockman <gdb@gregbrockman.com>
Co-authored-by: Christopher Hesse <christopherhesse@users.noreply.github.com>
2022-03-15 12:20:51 -07:00
Philippe Tillet
98ed7db8c1 [CODEGEN] Improvements and bugfixes (#463) 2022-02-24 14:56:24 -08:00
daadaada
a9dfdcaaa9 [FRONTEND] Make the performance model work for int8, tf32, and fp32 (#456) 2022-02-11 22:34:42 -08:00
Philippe Tillet
9b100302d3 [FRONTEND] Now using pybind11 to release GIL (#458) 2022-02-10 01:57:39 -08:00
Philippe Tillet
7b48340ffd [CI] Some fixes for the build (#451) 2022-02-06 19:11:33 -08:00
Philippe Tillet
5a8a544d10 [OPS][BLOCKSPARSE] Improved robustness, clarity and performance (#450)
* dds layout now internally re-uses dsd code path for increased code 
* at_mask and kp_mask related things are now dropped from the softmax API. I couldn't think of any case where it was needed beyond is_causal. And if there is any, we should probably find a way to get it implemented statically so that users don't have to materialize masks.
 * fixed bug in blocksparse matmul that caused troubles when layout had a full row/col of zeros
 * blocksparse softmax now no longer modifies any data in-place
 * blocksparse softmax now takes an is_dense arguments that provides better performance. Passing is_dense=True, is_causal=True is the best way to achieve triangular attention.
  * unit tests now test backward pass
2022-02-06 18:00:45 -08:00
TC
137bb67fad [LANG] Add fp16 to fp8 conversion (#444) 2022-02-02 20:42:09 -08:00
Philippe Tillet
b0d6e2f322 [STYLE] run autopep 2022-01-30 20:27:44 -08:00
Philippe Tillet
2922dc141c Merge branch 'master' into v2.0 2022-01-30 20:25:01 -08:00
Philippe Tillet
807d8a1945 [ALL] Merge master (#447) 2022-01-30 20:21:20 -08:00
Philippe Tillet
bef76b142a [BACKEND] float division is now approximate by default (#446) 2022-01-29 18:29:29 -08:00
Philippe Tillet
bd52e530a0 [OPS][BLOCKSPARSE] Fix padding issue in DSD LUT (#445) 2022-01-28 21:40:30 -08:00
daadaada
59d371c6eb [BACKEND] Added Int8 mma (#440) 2022-01-27 09:12:44 -08:00
Philippe Tillet
ccf9abe0ba [FRONTEND][RANDOM] Improved backward compatibility of RNG (#438)
The unsigned int PR definitely improved our RNG. However, it requires
different floating point arithmetics which, means the results are not
bit-wise identical to how they were before. This commit revives backward
compatibility, but we should change it back to the "right" way later.
2022-01-21 18:05:55 -08:00
Philippe Tillet
4c97d1ecd7 [FRONTEND] Bunch of fixes here and there (#436) 2022-01-20 10:55:59 -08:00
daadaada
2a944ded53 [TESTS] Added bfloat16 tests (#430) 2022-01-13 23:38:32 -08:00
Philippe Tillet
4c94359199 [FRONTEND] Alignment fix-up (#428) 2022-01-11 23:11:58 -08:00
Philippe Tillet
bbc78f6516 [FRONTEND][RANDOM] Make sure offset dtype is always uint32 before calling uint32_to_uniform_float (#427) 2022-01-11 11:08:49 -08:00
Botao Yu
bf32205edc [OPS][BLOCKSPARSE] Remove unnecessary loop and add cuda bool layout support (#425) 2022-01-11 11:07:16 -08:00
daadaada
94a2e10fe5 [BACKEND] Add bf16 & tf32 mma supports (on A100) (#426) 2022-01-11 10:20:31 -08:00
Madeleine Thompson
efdabe6073 [STYLE] check python with flake8 (#424)
I've been using this locally to find errors without running tests, and now that we're using autopep8, it passes with minimal suppressions. This is also what turned up the issues with the tutorials, which were fixed in #422.
2022-01-07 15:28:36 -08:00
Madeleine Thompson
a70acfec77 [STYLE] add isort and autopep8 config files and check on CI (#423)
Also a fix a few more style issues from the "aggressive" mode of autopep8.
2022-01-07 13:11:34 -08:00
Madeleine Thompson
9801aa7b56 [DOCS] fix tutorials for v2.0 (#422)
- Fix meta-parameter usage on tutorials.
- Install tutorial dependencies on CI.
- Switch from `requirements-test.txt` to `extras_require` for test dependencies, and also use it for tutorial dependencies.
- Make some performance tests deterministic.
2022-01-07 12:34:38 -08:00
Madeleine Thompson
8bf551ae7a [STYLE] run autopep8 and isort (#421)
Run:
```
isort ./python
autopep8 -i --ignore E501,E701,E731 $(find ./python/ -name '*.py')
```
with an `.isort.cfg` and then clean up a few warts. This PR should be a no-op; the idea is that this is all boring whitespace changes, and any config file changes will be in a different change to make it easier to review.
2022-01-06 14:34:17 -08:00