[WIP] Bulk executor initial implementation by ericl · Pull Request #30903 · ray-project/ray

ericl · 2022-12-05T22:38:38Z

Why are these changes needed?

Initial implementation of ray-project/enhancements#18

Original prototype: https://github.com/ray-project/ray/pull/30222/files

TODO:

Merge plan:

interfaces.py
memory debug utils
operator implementations with unit tests
bulk executor implementation with unit tests
legacy integration and flag for enabling
debug nightly test / regressions
enable by default

Signed-off-by: Eric Liang <ekhliang@gmail.com>

ericl · 2022-12-14T00:20:18Z

Some progress: now passing a good portion of dataset tests:

FAILED test_dataset.py::test_bulk_lazy_eval_split_mode[False] - AssertionError: (ObjectRef(00ffffffffffffffffffffffffffffffffffffff0100000001000000), BlockMetadata(num_rows=None, size_bytes=10, schema...
FAILED test_dataset.py::test_bulk_lazy_eval_split_mode[True] - AssertionError: (ObjectRef(00ffffffffffffffffffffffffffffffffffffff0100000009000000), BlockMetadata(num_rows=None, size_bytes=10, schema=...
FAILED test_dataset.py::test_basic_actors[False] - NotImplementedError
FAILED test_dataset.py::test_basic_actors[True] - NotImplementedError
FAILED test_dataset.py::test_callable_classes - NotImplementedError
FAILED test_dataset.py::test_convert_to_pyarrow - ModuleNotFoundError: No module named 'dask.dataframe'
FAILED test_dataset.py::test_iter_batches_local_shuffle[arrow-True] - ray.exceptions.RayTaskError: ray::_map_task() (pid=177449, ip=10.103.244.198)
FAILED test_dataset.py::test_iter_batches_local_shuffle[pandas-True] - ray.exceptions.ObjectFreedError: Failed to retrieve object 7f90397ff370237bffffffffffffffffffffffff0100000002000000. To see infor...
FAILED test_dataset.py::test_iter_batches_local_shuffle[simple-True] - ray.exceptions.RayTaskError: ray::_map_task() (pid=179727, ip=10.103.244.198)
FAILED test_dataset.py::test_map_batches_extra_args - ray.exceptions.RayTaskError(AssertionError): ray::_map_task() (pid=180287, ip=10.103.244.198)
FAILED test_dataset.py::test_map_batches_actors_preserves_order - NotImplementedError
FAILED test_dataset.py::test_map_batches_block_bundling_auto[1-2] - assert 10 == 5
FAILED test_dataset.py::test_map_batches_block_bundling_auto[1-3] - assert 10 == 4
FAILED test_dataset.py::test_map_batches_block_bundling_auto[1-4] - assert 10 == 3
FAILED test_dataset.py::test_map_batches_block_bundling_auto[2-4] - assert 10 == 5
FAILED test_dataset.py::test_map_batches_block_bundling_auto[1-5] - assert 10 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_auto[2-5] - assert 10 == 5
FAILED test_dataset.py::test_map_batches_block_bundling_auto[1-6] - assert 12 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_auto[2-6] - assert 10 == 4
FAILED test_dataset.py::test_map_batches_block_bundling_auto[3-6] - assert 10 == 5
FAILED test_dataset.py::test_map_batches_block_bundling_auto[1-7] - assert 14 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_auto[2-7] - assert 10 == 4
FAILED test_dataset.py::test_map_batches_block_bundling_auto[3-7] - assert 10 == 5
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_manual[block_sizes0-3-1] - assert 2 == 1
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_manual[block_sizes1-3-2] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_manual[block_sizes2-4-3] - assert 4 == 3
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_manual[block_sizes3-4-2] - assert 4 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_manual[block_sizes5-4-1] - assert 4 == 1
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_manual[block_sizes6-4-2] - assert 4 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes6-2] - assert 2 == 1
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes28-3] - assert 2 == 1
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes29-3] - assert 2 == 1
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes34-3] - assert 2 == 1
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes64-3] - assert 3 == 1
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes65-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes66-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes67-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes68-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes69-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes70-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes71-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes72-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes73-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes74-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes75-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes100-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes101-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes102-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes103-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes104-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes105-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes106-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes136-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes137-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes142-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes172-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes173-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes178-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes208-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes209-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes214-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes244-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes245-3] - assert 3 == 2
FAILED test_dataset.py::test_map_batches_block_bundling_skewed_auto[block_sizes250-3] - assert 3 == 2
FAILED test_dataset.py::test_from_dask - ModuleNotFoundError: No module named 'dask.dataframe'
FAILED test_dataset.py::test_to_dask[pandas] - ModuleNotFoundError: No module named 'dask.core'
FAILED test_dataset.py::test_to_dask[arrow] - ModuleNotFoundError: No module named 'dask.core'
FAILED test_dataset.py::test_to_dask_tensor_column_cast_pandas - ModuleNotFoundError: No module named 'dask.dataframe'
FAILED test_dataset.py::test_to_dask_tensor_column_cast_arrow - ModuleNotFoundError: No module named 'dask.dataframe'
FAILED test_dataset.py::test_from_modin - ImportError: cannot import name 'FilePathOrBuffer' from 'pandas._typing' (/home/eric/.local/lib/python3.8/site-packages/pandas/_typing.py)
FAILED test_dataset.py::test_to_modin - ImportError: cannot import name 'FilePathOrBuffer' from 'pandas._typing' (/home/eric/.local/lib/python3.8/site-packages/pandas/_typing.py)
FAILED test_dataset.py::test_map_batches_combine_empty_blocks - assert 30 == 3
FAILED test_dataset.py::test_random_shuffle[True-True] - ray.exceptions.RayTaskError: ray::_map_task() (pid=184295, ip=10.103.244.198)
FAILED test_dataset.py::test_random_shuffle[False-True] - ray.exceptions.RayTaskError: ray::_map_task() (pid=185254, ip=10.103.244.198)
FAILED test_dataset.py::test_dataset_retry_exceptions - AssertionError: (ObjectRef(00ffffffffffffffffffffffffffffffffffffff01000000ad6a0000), BlockMetadata(num_rows=None, size_bytes=20, schema=None, i...
FAILED test_dataset.py::test_split_is_not_disruptive - ray.exceptions.ObjectFreedError: Failed to retrieve object b6aef734ef5f822bffffffffffffffffffffffff0100000002000000. To see information about whe...
FAILED test_dataset.py::test_actor_pool_strategy_apply_interrupt - AssertionError: Legacy backend off
FAILED test_dataset.py::test_actor_pool_strategy_default_num_actors - NotImplementedError
================================================================== 78 failed, 539 passed, 3 skipped, 5010 warnings in 431.51s (0:07:11) ===================================================================

Signed-off-by: Eric Liang <ekhliang@gmail.com>

jianoaix · 2023-01-23T23:38:56Z

python/ray/data/_internal/execution/operators/actor_pool_submitter.py

            ray_remote_args: Remote arguments for the Ray actors to be created.
            pool_size: The size of the actor pool.
        """
-        if "num_cpus" not in ray_remote_args:


@ericl I reverted this, since it looks not compatible with the requirement that "num_gpus and num_cpus cannot be both specified" (user may want to run actors on gpus and then they won't be able tp specify cpus).

…executorimpl

jianoaix · 2023-01-24T23:31:22Z

All CI tests are passing now. However, we seem to have a release test failure that may be relevant: https://buildkite.com/ray-project/release-tests-pr/builds/26265#0185e158-7979-4c94-be61-22da62d208ca

jianoaix · 2023-01-24T23:40:54Z

All CI tests are passing now. However, we seem to have a release test failure that may be relevant: https://buildkite.com/ray-project/release-tests-pr/builds/26265#0185e158-7979-4c94-be61-22da62d208ca

It seems likely due to the lack of autoscaling for actor pool, as it had just one actor through the entire run, @clarkzinzow

And since it's about the same issue as TODO in unit test, shall we note this as a debt to fix soon and merge this PR? @ericl

2023-01-23 20:47:03,018	INFO bulk_executor.py:39 -- Executing DAG InputDataBuffer[Input] -> MapOperator[map_batches]
--
  |  
  | 0%\|          \| 0/363 [00:00<?, ?it/s]
  | map_batches:   0%\|          \| 0/363 [00:00<?, ?it/s]
  | map_batches:   0%\|          \| 1/363 [00:18<1:52:33, 18.66s/it]
  | map_batches, 1 actors:   0%\|          \| 1/363 [00:18<1:52:33, 18.66s/it](raylet, ip=172.31.203.20) Spilled 34054 MiB, 110 objects, write throughput 637 MiB/s.
  |  
  | map_batches, 1 actors:   1%\|          \| 2/363 [00:37<1:54:07, 18.97s/it]
  | map_batches, 1 actors:   1%\|          \| 3/363 [01:28<3:21:07, 33.52s/it]
  | map_batches, 1 actors:   1%\|          \| 4/363 [01:42<2:35:06, 25.92s/it]
  | map_batches, 1 actors:   1%\|▏         \| 5/363 [01:44<1:42:34, 17.19s/it]
  | map_batches, 1 actors:   2%\|▏         \| 6/363 [01:45<1:10:09, 11.79s/it]
  | map_batches, 1 actors:   2%\|▏         \| 7/363 [01:48<51:34,  8.69s/it]
  | map_batches, 1 actors:   2%\|▏         \| 8/363 [01:49<37:42,  6.37s/it]
  | map_batches, 1 actors:   2%\|▏         \| 9/363 [01:51<28:59,  4.91s/it]
  | map_batches, 1 actors:   3%\|▎         \| 10/363 [01:52<22:21,  3.80s/it]
  | map_batches, 1 actors:   3%\|▎         \| 11/363 [01:54<19:05,  3.25s/it]
  | map_batches, 1 actors:   3%\|▎         \| 12/363 [01:56<15:44,  2.69s/it]
  | map_batches, 1 actors:   4%\|▎         \| 13/363 [01:58<14:29,  2.49s/it]
  | map_batches, 1 actors:   4%\|▍         \| 14/363 [01:59<12:23,  2.13s/it]
  | map_batches, 1 actors:   4%\|▍         \| 15/363 [02:01<11:58,  2.06s/it]
  | map_batches, 1 actors:   4%\|▍         \| 16/363 [02:02<10:36,  1.84s/it]
  | map_batches, 1 actors:   5%\|▍         \| 17/363 [02:04<10:42,  1.86s/it]
  | map_batches, 1 actors:   5%\|▍         \| 18/363 [02:05<09:43,  1.69s/it]
  | map_batches, 1 actors:   5%\|▌         \| 19/363 [02:07<10:04,  1.76s/it]
  | map_batches, 1 actors:   6%\|▌         \| 20/363 [02:09<09:26,  1.65s/it]
  | map_batches, 1 actors:   6%\|▌         \| 21/363 [02:11<09:51,  1.73s/it]
  | map_batches, 1 actors:   6%\|▌         \| 22/363 [02:12<09:17,  1.64s/it]
  | map_batches, 1 actors:   6%\|▋         \| 23/363 [02:14<09:54,  1.75s/it]
  | map_batches, 1 actors:   7%\|▋         \| 24/363 [02:15<09:07,  1.62s/it]
  | map_batches, 1 actors:   7%\|▋         \| 25/363 [02:17<08:55,  1.58s/it]
  | map_batches, 1 actors:   7%\|▋         \| 26/363 [02:18<08:56,  1.59s/it]
  | map_batches, 1 actors:   7%\|▋         \| 27/363 [02:20<08:05,  1.45s/it]
  | map_batches, 1 actors:   8%\|▊         \| 28/363 [02:21<08:40,  1.55s/it]
  | map_batches, 1 actors:   8%\|▊         \| 29/363 [02:23<08:14,  1.48s/it]
  | map_batches, 1 actors:   8%\|▊         \| 30/363 [02:25<08:55,  1.61s/it]
  | map_batches, 1 actors:   9%\|▊         \| 31/363 [02:26<08:24,  1.52s/it]
  | map_batches, 1 actors:   9%\|▉         \| 32/363 [02:28<09:01,  1.64s/it]
  | map_batches, 1 actors:   9%\|▉         \| 33/363 [02:29<08:17,  1.51s/it]
  | map_batches, 1 actors:   9%\|▉         \| 34/363 [02:30<08:05,  1.48s/it]
  ......
  | map_batches, 1 actors:  90%\|████████▉ \| 325/363 [09:53<00:43,  1.15s/it]
  | map_batches, 1 actors:  90%\|████████▉ \| 326/363 [09:54<00:38,  1.05s/it]
  | map_batches, 1 actors:  90%\|█████████ \| 327/363 [09:55<00:36,  1.01s/it]
  | map_batches, 1 actors:  90%\|█████████ \| 328/363 [09:56<00:37,  1.06s/it]
  | map_batches, 1 actors:  91%\|█████████ \| 329/363 [09:57<00:37,  1.11s/it]
  | map_batches, 1 actors:  91%\|█████████ \| 330/363 [09:58<00:33,  1.02s/it]
  | map_batches, 1 actors:  91%\|█████████ \| 331/363 [09:59<00:30,  1.05it/s]
  | map_batches, 1 actors:  91%\|█████████▏\| 332/363 [10:00<00:35,  1.15s/it]
  | map_batches, 1 actors:  92%\|█████████▏\| 333/363 [10:01<00:31,  1.04s/it]
  | map_batches, 1 actors:  92%\|█████████▏\| 334/363 [10:02<00:29,  1.00s/it]
  | map_batches, 1 actors:  92%\|█████████▏\| 335/363 [10:04<00:33,  1.18s/it]
  | map_batches, 1 actors:  93%\|█████████▎\| 336/363 [10:05<00:28,  1.07s/it]
  | map_batches, 1 actors:  93%\|█████████▎\| 337/363 [10:05<00:25,  1.01it/s]
  | map_batches, 1 actors:  93%\|█████████▎\| 338/363 [10:06<00:24,  1.04it/s]
  | map_batches, 1 actors:  93%\|█████████▎\| 339/363 [10:08<00:27,  1.13s/it]
  | map_batches, 1 actors:  94%\|█████████▎\| 340/363 [10:09<00:23,  1.03s/it]
  | map_batches, 1 actors:  94%\|█████████▍\| 341/363 [10:09<00:21,  1.01it/s]
  | map_batches, 1 actors:  94%\|█████████▍\| 342/363 [10:11<00:25,  1.20s/it]
  | map_batches, 1 actors:  94%\|█████████▍\| 343/363 [10:12<00:21,  1.08s/it]
  | map_batches, 1 actors:  95%\|█████████▍\| 344/363 [10:13<00:19,  1.03s/it]
  | map_batches, 1 actors:  95%\|█████████▌\| 345/363 [10:15<00:21,  1.20s/it]
  | map_batches, 1 actors:  95%\|█████████▌\| 346/363 [10:15<00:18,  1.08s/it]
  | map_batches, 1 actors:  96%\|█████████▌\| 347/363 [10:16<00:15,  1.00it/s]
  | map_batches, 1 actors:  96%\|█████████▌\| 348/363 [10:17<00:14,  1.03it/s]
  | map_batches, 1 actors:  96%\|█████████▌\| 349/363 [10:19<00:16,  1.16s/it]
  | map_batches, 1 actors:  96%\|█████████▋\| 350/363 [10:19<00:13,  1.05s/it]
  | map_batches, 1 actors:  97%\|█████████▋\| 351/363 [10:20<00:11,  1.02it/s]
  | map_batches, 1 actors:  97%\|█████████▋\| 352/363 [10:22<00:13,  1.20s/it]
  | map_batches, 1 actors:  97%\|█████████▋\| 353/363 [10:23<00:10,  1.08s/it]
  | map_batches, 1 actors:  98%\|█████████▊\| 354/363 [10:24<00:08,  1.01it/s]
  | map_batches, 1 actors:  98%\|█████████▊\| 355/363 [10:25<00:08,  1.03s/it]
  | map_batches, 1 actors:  98%\|█████████▊\| 356/363 [10:26<00:07,  1.11s/it]
  | map_batches, 1 actors:  98%\|█████████▊\| 357/363 [10:27<00:06,  1.02s/it]
  | map_batches, 1 actors:  99%\|█████████▊\| 358/363 [10:28<00:04,  1.02it/s]
  | map_batches, 1 actors:  99%\|█████████▉\| 359/363 [10:29<00:04,  1.17s/it]
  | map_batches, 1 actors:  99%\|█████████▉\| 360/363 [10:30<00:03,  1.06s/it]
  | map_batches, 1 actors:  99%\|█████████▉\| 361/363 [10:31<00:02,  1.01s/it]
  | map_batches, 1 actors: 100%\|█████████▉\| 362/363 [10:32<00:01,  1.16s/it]
  | map_batches, 1 actors: 100%\|██████████\| 363/363 [10:33<00:00,  1.04s/it]
  | map_batches, 0 actors: 100%\|██████████\| 363/363 [10:33<00:00,  1.04s/it]
  | map_batches, 0 actors: 100%\|██████████\| 363/363 [10:33<00:00,  1.75s/it]
  | run_xgboost_prediction takes 671.7315916329999 seconds.
  | Results: {'training_time': 781.2928511609999, 'prediction_time': 671.7315916329999}
  | Traceback (most recent call last):
  | File "workloads/xgboost_benchmark.py", line 175, in <module>
  | main(args)
  | File "workloads/xgboost_benchmark.py", line 155, in main
  | f"Batch prediction on XGBoost is taking {prediction_time} seconds, "
  | RuntimeError: Batch prediction on XGBoost is taking 671.7315916329999 seconds, which is longer than expected (450 seconds).

ericl · 2023-01-25T00:17:31Z

Alright. Let's leave the TODO to fix this test (or we can fix this test by increasing the min pool size).

…-project#31283) Add a utility class for tracing object allocation / freeing. This makes it a lot easier to debug memory allocation / freeing issues. This is split out from ray-project#30903 Signed-off-by: tmynn <hovhannes.tamoyan@gmail.com>

…oject#31305) Add the initial operator implementations. This is split out from ray-project#30903 Signed-off-by: tmynn <hovhannes.tamoyan@gmail.com>

…oject#31443) Add the basic bulk executor. This is split out from ray-project#30903 Signed-off-by: tmynn <hovhannes.tamoyan@gmail.com>

ericl added 7 commits December 5, 2022 14:31

copy prototype

9e4451e

cleanup

8924a89

wip compatibility

44578ce

Signed-off-by: Eric Liang <ekhliang@gmail.com>

add basic wiring

e0a346a

Signed-off-by: Eric Liang <ekhliang@gmail.com>

works

22504c0

Signed-off-by: Eric Liang <ekhliang@gmail.com>

fix up split handling

0b26570

Signed-off-by: Eric Liang <ekhliang@gmail.com>

refactor legacy compat package

3f0e0cb

Signed-off-by: Eric Liang <ekhliang@gmail.com>

ericl force-pushed the bulk-executor branch from bac87d6 to 3f0e0cb Compare December 6, 2022 22:57

ericl added 3 commits December 6, 2022 16:27

todo move operators fully

eaa46b0

Signed-off-by: Eric Liang <ekhliang@gmail.com>

reorganize opeators

3162f44

stub out actors impl

2136170

ericl force-pushed the bulk-executor branch from 94a64b9 to 2136170 Compare December 7, 2022 01:44

ericl added 3 commits December 6, 2022 17:56

improve legacy integration

38ae324

add str

9f24555

add own block propagation

f33c772

Signed-off-by: Eric Liang <ekhliang@gmail.com>

ericl assigned ericl and unassigned ericl Dec 7, 2022

ericl added 2 commits December 6, 2022 22:04

rename to tasks

bf5288f

Signed-off-by: Eric Liang <ekhliang@gmail.com>

add basic stats

f5efe2c

ericl force-pushed the bulk-executor branch from 4ec0bc4 to f5efe2c Compare December 13, 2022 21:01

ericl added 2 commits December 13, 2022 15:56

implement alltoall

e5790dc

Signed-off-by: Eric Liang <ekhliang@gmail.com>

revert format change

5c7e490

Signed-off-by: Eric Liang <ekhliang@gmail.com>

ericl added 7 commits December 13, 2022 16:22

Merge remote-tracking branch 'upstream/master' into bulk-executor

d6bee3c

fixme

1eb5519

Signed-off-by: Eric Liang <ekhliang@gmail.com>

fix

ec66fd0

Signed-off-by: Eric Liang <ekhliang@gmail.com>

fix own propagation

5aa082b

add debug mem metrics

c8f8c79

Signed-off-by: Eric Liang <ekhliang@gmail.com>

fix block clearing for datasetpipeline

5b2f7ec

add config

00025f5

Signed-off-by: Eric Liang <ekhliang@gmail.com>

jianoaix reviewed Jan 23, 2023

View reviewed changes

jianoaix added 2 commits January 24, 2023 00:43

fix bazel test

06b1ad7

Merge branch 'master' of https://github.com/ray-project/ray into bulk…

3867061

…executorimpl

jianoaix mentioned this pull request Jan 24, 2023

Fixes for the new bulk execution backend #31884

Merged

7 tasks

jianoaix added 3 commits January 24, 2023 19:38

Merge branch 'master' of https://github.com/ray-project/ray into bulk…

7097973

…executorimpl

minimize dif

0b74edf

less diff

a9a66ab

jianoaix marked this pull request as ready for review January 24, 2023 19:43

jianoaix requested review from c21, jjyao and scv119 as code owners January 24, 2023 19:43

disable incremental take test

a265437

jianoaix approved these changes Jan 24, 2023

View reviewed changes

ericl merged commit 877770e into ray-project:master Jan 25, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[WIP] Bulk executor initial implementation#30903

[WIP] Bulk executor initial implementation#30903
ericl merged 152 commits intoray-project:masterfrom
ericl:bulk-executor

ericl commented Dec 5, 2022 •

edited

Loading

Uh oh!

ericl commented Dec 14, 2022

Uh oh!

jianoaix Jan 23, 2023

Uh oh!

jianoaix commented Jan 24, 2023

Uh oh!

jianoaix commented Jan 24, 2023 •

edited

Loading

Uh oh!

ericl commented Jan 25, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

ericl commented Dec 5, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Why are these changes needed?

Uh oh!

ericl commented Dec 14, 2022

Uh oh!

jianoaix Jan 23, 2023

Choose a reason for hiding this comment

Uh oh!

jianoaix commented Jan 24, 2023

Uh oh!

jianoaix commented Jan 24, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ericl commented Jan 25, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ericl commented Dec 5, 2022 •

edited

Loading

jianoaix commented Jan 24, 2023 •

edited

Loading