Chirag Pandya 485497e727 [c10d][fr] flight recorder improvements (#143446) ... Summary: 1. Flight recorder dumps are now automatically dumped by default upon timeout or exception. Users don't need to opt-in. 2. Change default dump location to running user's home directory `.cache` folder. Test Plan: 1. Tested locally by running the crash program from flight recorder tutorial page. https://pytorch.org/tutorials/prototype/flight_recorder_tutorial.html#an-end-to-end-example 2. Noted that flight recorder files were correctly created. ❯ pwd /home/cpio/.cache/fr_trace ❯ ls nccl_trace_rank_0 nccl_trace_rank_1 Differential Revision: [D67363720](https://our.internmc.facebook.com/intern/diff/D67363720) Pull Request resolved: https://github.com/pytorch/pytorch/pull/143446 Approved by: https://github.com/d4l3k		2024-12-20 20:41:30 +00:00
..
_composable	[FSDP2] Clamp reduce_dtype in lazy init (#143297)	2024-12-17 00:25:08 +00:00
_shard	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
_tensor	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
_tools	[FSDP2] Move to public torch.distributed.fsdp (#141868)	2024-12-07 01:24:28 +00:00
algorithms	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
bin
checkpoint	[state dict] Change _load_model_state_dict to enable cpu_offload, accept 2 device type and optimize memory (#142845)	2024-12-19 05:06:41 +00:00
elastic	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
flight_recorder	[fr] recognize all_reduce_barrier as a valid op (#143354)	2024-12-17 21:09:18 +00:00
fsdp	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
launcher	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
nn/jit
optim
pipelining	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
rpc
tensor/parallel	[fused_all_gather_matmul] introduce an argument to specify whether the all-gather result needs to be returned (#143159)	2024-12-17 01:07:27 +00:00
argparse_util_test.py
test_backends.py
test_c10d_common.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_c10d_functional_native.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_c10d_gloo.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_c10d_logger.py
test_c10d_nccl.py	[c10d][fr] flight recorder improvements (#143446)	2024-12-20 20:41:30 +00:00
test_c10d_object_collectives.py
test_c10d_ops_nccl.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_c10d_pypg.py	DistributedDataParallel: add init_sync option to control collectives during initialization (#142824)	2024-12-11 20:28:38 +00:00
test_c10d_spawn.py
test_c10d_spawn_gloo.py
test_c10d_spawn_nccl.py
test_c10d_spawn_ucc.py
test_c10d_ucc.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_collective_utils.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_compute_comm_reordering.py
test_control_collectives.py
test_data_parallel.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_device_mesh.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_distributed_spawn.py
test_dynamo_distributed.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_fake_pg.py
test_functional_api.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_inductor_collectives.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_launcher.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_multi_threaded_pg.py
test_nccl.py
test_pg_wrapper.py
test_store.py	Fix unused Python variables in test/[a-d]* (#134665)	2024-12-13 22:13:12 +00:00
test_symmetric_memory.py	[fused_all_gather_matmul] use _multimem_all_gather_matmul for small global Ms (#143160)	2024-12-17 01:07:27 +00:00