wehub-resource-sync
49 lines
2.0 KiB
Python
49 lines
2.0 KiB
Python
# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import legacy_test.test_collective_api_base as test_base
|
|
|
|
import paddle
|
|
from paddle import base
|
|
|
|
|
|
class TestCollectiveAllgatherObjectAPI(test_base.TestCollectiveAPIRunnerBase):
|
|
def __init__(self):
|
|
self.global_ring_id = 0
|
|
|
|
def get_model(self, main_prog, startup_program, rank, indata=None):
|
|
with base.program_guard(main_prog, startup_program):
|
|
# Run the collective twice, once per supported initialization
|
|
# style, and assert the results match. This locks in the
|
|
# alignment with torch.distributed.all_gather_object: both an
|
|
# empty list (Paddle legacy) and a pre-allocated [None]*world_size
|
|
# (PyTorch) must produce identical output. Doubles the comm cost
|
|
# for this test but stays well inside the 120s timeout.
|
|
paddle_style = []
|
|
paddle.distributed.all_gather_object(paddle_style, indata)
|
|
|
|
world_size = paddle.distributed.get_world_size()
|
|
torch_style = [None for _ in range(world_size)]
|
|
paddle.distributed.all_gather_object(torch_style, indata)
|
|
|
|
assert paddle_style == torch_style, (
|
|
f"all_gather_object initialization styles disagree: "
|
|
f"empty-list {paddle_style!r} vs pre-allocated {torch_style!r}"
|
|
)
|
|
return torch_style
|
|
|
|
|
|
if __name__ == "__main__":
|
|
test_base.runtime_main(TestCollectiveAllgatherObjectAPI, "allgather_object")
|