Skip to content

[RPC tests] Merge tests for faulty agent into single script #40817

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 13 commits into from
Closed

[RPC tests] Merge tests for faulty agent into single script #40817

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
d89fc7b
[RPC tests] Merge tests for faulty agent into single script
lw Jun 30, 2020
e0b1533
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 1, 2020
d85fd92
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 1, 2020
9fa25e1
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 1, 2020
76ecc85
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 1, 2020
a4a91c0
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 2, 2020
67a3b42
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 3, 2020
e8aec6b
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 10, 2020
f521032
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 16, 2020
f949c9f
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Jul 21, 2020
dd37719
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Aug 4, 2020
b15fb6f
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Aug 5, 2020
05cc4f8
Update on "[RPC tests] Merge tests for faulty agent into single script"
lw Aug 5, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 0 additions & 21 deletions test/distributed/rpc/faulty_agent/test_dist_autograd_spawn.py
View file
Open in desktop

This file was deleted.

19 changes: 0 additions & 19 deletions test/distributed/rpc/faulty_agent/test_rpc_spawn.py
View file
Open in desktop

This file was deleted.

21 changes: 0 additions & 21 deletions test/distributed/rpc/jit/faulty_agent/test_rpc_spawn.py
View file
Open in desktop

This file was deleted.

46 changes: 46 additions & 0 deletions test/distributed/rpc/test_faulty_agent.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,46 @@
#!/usr/bin/env python3
import unittest

from torch.testing._internal.common_distributed import MultiProcessTestCase
Comment on lines +1 to +4
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Similar to my comment in #40816 (comment), is it possible to avoid having separate files and enable different agents based on env variables?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

See my comment there.

from torch.testing._internal.common_utils import TEST_WITH_ASAN, run_tests
from torch.testing._internal.distributed.rpc.dist_autograd_test import (
FaultyAgentDistAutogradTest,
)
from torch.testing._internal.distributed.rpc.faulty_rpc_agent_test_fixture import (
FaultyRpcAgentTestFixture,
)
from torch.testing._internal.distributed.rpc.jit.rpc_test_faulty import (
JitFaultyAgentRpcTest,
)
from torch.testing._internal.distributed.rpc.rpc_test import FaultyAgentRpcTest


@unittest.skipIf(
TEST_WITH_ASAN, "Skip ASAN as torch + multiprocessing spawn have known issues"
)
class SpawnHelper(MultiProcessTestCase):
def setUp(self):
super().setUp()
self._spawn_processes()


class FaultyRpcTestWithSpawn(
FaultyRpcAgentTestFixture, FaultyAgentRpcTest, SpawnHelper
):
pass


class FaultyDistAutogradTestWithSpawn(
FaultyRpcAgentTestFixture, FaultyAgentDistAutogradTest, SpawnHelper
):
pass


class FaultyJitRpcTestWithSpawn(
FaultyRpcAgentTestFixture, JitFaultyAgentRpcTest, SpawnHelper
):
pass


if __name__ == "__main__":
run_tests()
10 changes: 3 additions & 7 deletions test/run_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -78,8 +78,7 @@
'test_profiler',
'distributed/nn/jit/test_instantiator',
'distributed/nn/api/test_remote_module_spawn',
'distributed/rpc/faulty_agent/test_dist_autograd_spawn',
'distributed/rpc/faulty_agent/test_rpc_spawn',
'distributed/rpc/test_faulty_agent',
'distributed/rpc/jit/test_dist_autograd_spawn',
'distributed/rpc/test_tensorpipe_agent',
'distributed/rpc/test_dist_autograd_spawn',
Expand All @@ -88,16 +87,14 @@
'test_jit_py3',
'test_determination',
'distributed/rpc/jit/test_rpc_spawn',
'distributed/rpc/faulty_agent/test_rpc_spawn',
'test_futures',
'distributed/test_ddp_under_dist_autograd',
]

WINDOWS_BLOCKLIST = [
'distributed/nn/jit/test_instantiator',
'distributed/nn/api/test_remote_module_spawn',
'distributed/rpc/faulty_agent/test_dist_autograd_spawn',
'distributed/rpc/faulty_agent/test_rpc_spawn',
'distributed/rpc/test_faulty_agent',
'distributed/rpc/jit/test_dist_autograd_spawn',
'distributed/rpc/jit/test_rpc_spawn',
'distributed/rpc/test_tensorpipe_agent',
Expand All @@ -111,8 +108,7 @@
ROCM_BLOCKLIST = [
'distributed/nn/jit/test_instantiator',
'distributed/nn/api/test_remote_module_spawn',
'distributed/rpc/faulty_agent/test_dist_autograd_spawn',
'distributed/rpc/faulty_agent/test_rpc_spawn',
'distributed/rpc/test_faulty_agent',
'distributed/rpc/jit/test_dist_autograd_spawn',
'distributed/rpc/jit/test_rpc_spawn',
'distributed/rpc/test_tensorpipe_agent',
Expand Down
5 changes: 1 addition & 4 deletions torch/testing/_internal/distributed/rpc/dist_autograd_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,9 +23,6 @@
from torch.testing._internal.distributed.rpc.rpc_agent_test_fixture import (
RpcAgentTestFixture,
)
from torch.testing._internal.distributed.rpc.faulty_rpc_agent_test_fixture import (
FaultyRpcAgentTestFixture,
)
from torch.testing._internal.common_distributed import skip_if_lt_x_gpu


Expand Down Expand Up @@ -2198,7 +2195,7 @@ def test_gpu_to_cpu_continuation_gpu_root(self):
local_grads = ret if ret else local_grads


class FaultyAgentDistAutogradTest(FaultyRpcAgentTestFixture):
class FaultyAgentDistAutogradTest(RpcAgentTestFixture):
# Reusing a simplified helper function from DistAutogradTest to ensure
# autograd context is successfully cleaned up even when RPCs are failing.
def context_cleanup_test_helper(self, rpc_args, func):
Expand Down
6 changes: 3 additions & 3 deletions torch/testing/_internal/distributed/rpc/jit/rpc_test_faulty.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,8 +10,8 @@
worker_name,
wait_until_pending_futures_and_users_flushed
)
from torch.testing._internal.distributed.rpc.faulty_rpc_agent_test_fixture import (
FaultyRpcAgentTestFixture,
from torch.testing._internal.distributed.rpc.rpc_agent_test_fixture import (
RpcAgentTestFixture,
)


Expand Down Expand Up @@ -82,7 +82,7 @@ def rpc_async_with_rref_arg(dst_worker_name, args):
return ret


class JitFaultyAgentRpcTest(FaultyRpcAgentTestFixture):
class JitFaultyAgentRpcTest(RpcAgentTestFixture):
"""
Run tests for rpc_async in JIT under the faulty agent test fixture to test
arbitrary timeouts.
Expand Down
5 changes: 1 addition & 4 deletions torch/testing/_internal/distributed/rpc/rpc_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,9 +39,6 @@
RpcAgentTestFixture,
)
from torch.testing._internal.common_utils import TemporaryFileName
from torch.testing._internal.distributed.rpc.faulty_rpc_agent_test_fixture import (
FaultyRpcAgentTestFixture,
)


def foo_add():
Expand Down Expand Up @@ -3432,7 +3429,7 @@ def test_wait_all_with_partial_exception(self):
ret = torch.futures.wait_all(futs)


class FaultyAgentRpcTest(FaultyRpcAgentTestFixture):
class FaultyAgentRpcTest(RpcAgentTestFixture):

# no faulty_messages defined so this fails all retryable messages - see
# faulty_rpc_agent_test_fixture.py for the list of retryable messages.
Expand Down