pytorch/test/distributed
Wanchao Liang 848cfe8d45 [reland] unflatten_tensor on compute stream for DTensorExtension (#117020)
reland of https://github.com/pytorch/pytorch/pull/116559, which was reverted by internal.

The underlying reason for the revert is that the torch.dynamo.disable can't be used by the
pytorch codebase, as it's conflicting with some torch.deploy together, although the later one
only run some inference, but it somehow take that weird dependency on fsdp..

We have seen this issue with our functional collectives that we can't
use any dynamo components otherwise torch.deploy would complain..

verified internally that after removing torch.dynamo.disable the test
passed again

Pull Request resolved: https://github.com/pytorch/pytorch/pull/117020
Approved by: https://github.com/awgu
2024-01-09 21:25:15 +00:00
..
_composable [DCP][BE] Move DCP._state_dict_utils out from DCP (#115523) 2023-12-13 08:59:48 +00:00
_shard Refactor some tests by using TEST_CUDA & TEST_MULTIGPU instead (#116083) 2024-01-03 08:53:59 +00:00
_spmd [DeviceMesh] Rename get_dim_groups to get_group (#114708) 2023-11-30 23:40:14 +00:00
_tensor [dtensor] fix dtensor _to_copy op for mix precision (#116426) 2024-01-03 07:29:08 +00:00
_tools
algorithms
bin
checkpoint [BE]: Enable F821 and fix bugs (#116579) 2024-01-01 08:40:46 +00:00
elastic [BE]: Enable F821 and fix bugs (#116579) 2024-01-01 08:40:46 +00:00
fsdp [reland] unflatten_tensor on compute stream for DTensorExtension (#117020) 2024-01-09 21:25:15 +00:00
launcher
nn/jit
optim
pipeline/sync Refactor some tests by using TEST_CUDA & TEST_MULTIGPU instead (#116083) 2024-01-03 08:53:59 +00:00
rpc [BE]: Enable F821 and fix bugs (#116579) 2024-01-01 08:40:46 +00:00
tensor/parallel [tp] disable some assertion temporarily for torch.compile (#116573) 2024-01-03 23:01:19 +00:00
argparse_util_test.py
test_c10d_common.py [c10d] Fix the hang issue in store.check(TIMEOUT_DUMP) (#116297) 2023-12-22 04:04:30 +00:00
test_c10d_functional_native.py [AOTI] Add pybind for AOTIModelContainerRunnerCpu and AOTIModelContainerRunnerCuda (#116269) 2024-01-04 18:58:24 +00:00
test_c10d_gloo.py [c10d] Create a python c10d API _set_pg_timeout to set timeout (#115453) 2023-12-12 20:52:43 +00:00
test_c10d_logger.py
test_c10d_nccl.py Refactor some tests by using TEST_CUDA & TEST_MULTIGPU instead (#116083) 2024-01-03 08:53:59 +00:00
test_c10d_object_collectives.py
test_c10d_pypg.py
test_c10d_spawn.py [BE]: Update flake8 to v6.1.0 and fix lints (#116591) 2024-01-03 06:04:44 +00:00
test_c10d_spawn_gloo.py
test_c10d_spawn_nccl.py
test_c10d_spawn_ucc.py
test_c10d_ucc.py
test_collective_utils.py
test_compute_comm_reordering.py
test_data_parallel.py
test_device_mesh.py Make input contiguous for DTensor reduce scatter to fix the incorrect numerical values (#115847) 2023-12-17 01:35:09 +00:00
test_distributed_spawn.py
test_dynamo_distributed.py [dynamo / DDP] Add optimize_ddp_lazy_compile config to control lazy compile for DDPOptimizer (False by default) (#116292) 2023-12-21 22:34:24 +00:00
test_fake_pg.py further deprecate PairwiseParallel and SequenceParallel from test (#114402) 2023-11-30 05:06:08 +00:00
test_functional_api.py follow up #115078, broken distributed tests (#116217) 2024-01-08 15:26:54 +00:00
test_inductor_collectives.py [dynamo][collective_op] Use the value of the wrappered attribute async_op in dynamo when checking supported or not (#115921) 2023-12-21 03:27:57 +00:00
test_launcher.py
test_multi_threaded_pg.py
test_nccl.py [BE]: Enable F821 and fix bugs (#116579) 2024-01-01 08:40:46 +00:00
test_pg_wrapper.py Switch env variable use in test harnesses to the non-deprecated names to fix warnings (#114880) 2023-12-01 20:08:23 +00:00
test_store.py [BE]: Enable F821 and fix bugs (#116579) 2024-01-01 08:40:46 +00:00