Skip to content
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.

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

fix a100 cuda 12 ut #54542

Merged
merged 2 commits into from
Jun 12, 2023
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
16 changes: 12 additions & 4 deletions test/collective/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -200,8 +200,12 @@ if((WITH_GPU OR WITH_ROCM) AND (LINUX))
endif()
if((WITH_GPU OR WITH_ROCM) AND (LINUX))
py_test_modules(
test_collective_reduce_api MODULES test_collective_reduce_api ENVS
"http_proxy=;https_proxy=;PYTHONPATH=..:${PADDLE_BINARY_DIR}/python")
test_collective_reduce_api
MODULES
test_collective_reduce_api
ENVS
"NVIDIA_TF32_OVERRIDE=0;http_proxy=;https_proxy=;PYTHONPATH=..:${PADDLE_BINARY_DIR}/python"
)
set_tests_properties(test_collective_reduce_api
PROPERTIES TIMEOUT "500" LABELS "RUN_TYPE=DIST")
endif()
Expand Down Expand Up @@ -272,8 +276,12 @@ if((WITH_GPU OR WITH_ROCM) AND (LINUX))
endif()
if((WITH_GPU OR WITH_ROCM) AND (LINUX))
py_test_modules(
test_collective_split_col_linear MODULES test_collective_split_col_linear
ENVS "http_proxy=;https_proxy=;PYTHONPATH=..:${PADDLE_BINARY_DIR}/python")
test_collective_split_col_linear
MODULES
test_collective_split_col_linear
ENVS
"NVIDIA_TF32_OVERRIDE=0;http_proxy=;https_proxy=;PYTHONPATH=..:${PADDLE_BINARY_DIR}/python"
)
set_tests_properties(test_collective_split_col_linear
PROPERTIES TIMEOUT "300" LABELS "RUN_TYPE=DIST")
endif()
Expand Down
19 changes: 13 additions & 6 deletions test/collective/fleet/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -332,7 +332,7 @@ if(LOCAL_ALL_ARCH AND LOCAL_ALL_PLAT)
LABELS
"RUN_TYPE=DIST"
ENVS
"PADDLE_DIST_UT_PORT=21234;http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python"
"NVIDIA_TF32_OVERRIDE=0;PADDLE_DIST_UT_PORT=21234;http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python"
)
set_tests_properties(test_hybrid_parallel_inference_helper PROPERTIES TIMEOUT
"120")
Expand All @@ -351,8 +351,12 @@ if((WITH_GPU OR WITH_ROCM) AND (LINUX OR WIN32))
endif()
if(LOCAL_ALL_ARCH AND LOCAL_ALL_PLAT)
py_test_modules(
test_recv_save_op MODULES test_recv_save_op ENVS
"http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python")
test_recv_save_op
MODULES
test_recv_save_op
ENVS
"NVIDIA_TF32_OVERRIDE=0;http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python"
)
endif()
if(LOCAL_ALL_ARCH AND LOCAL_ALL_PLAT)
py_test_modules(
Expand Down Expand Up @@ -696,7 +700,7 @@ if((WITH_GPU OR WITH_ROCM) AND LOCAL_ALL_PLAT)
LABELS
"RUN_TYPE=DIST"
ENVS
"PADDLE_DIST_UT_PORT=21274;http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python"
"NVIDIA_TF32_OVERRIDE=0;PADDLE_DIST_UT_PORT=21274;http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python"
)
set_tests_properties(test_parallel_dygraph_mnist PROPERTIES TIMEOUT "200")
endif()
Expand Down Expand Up @@ -922,9 +926,12 @@ if((WITH_GPU) AND (LINUX))
endif()
if((WITH_GPU) AND (LINUX))
py_test_modules(
test_dygraph_save_for_auto_infer MODULES test_dygraph_save_for_auto_infer
test_dygraph_save_for_auto_infer
MODULES
test_dygraph_save_for_auto_infer
ENVS
"http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python")
"NVIDIA_TF32_OVERRIDE=0;http_proxy=;https_proxy=;PYTHONPATH=../..:${PADDLE_BINARY_DIR}/python"
)
set_tests_properties(test_dygraph_save_for_auto_infer
PROPERTIES TIMEOUT "300" LABELS "RUN_TYPE=DIST")
endif()
2 changes: 1 addition & 1 deletion test/collective/fleet/hybrid_parallel_mp_layers.py
Original file line number Diff line number Diff line change
Expand Up @@ -221,7 +221,7 @@ def test_row_parallel_layer(self):
optimizer_b.step()

np.testing.assert_allclose(
loss_a.numpy(), loss_b.numpy(), rtol=5e-6
loss_a.numpy(), loss_b.numpy(), rtol=5e-5
)

def test_parallel_embedding(self):
Expand Down
2 changes: 1 addition & 1 deletion test/distributed_passes/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,7 @@ if(NOT ((WITH_GPU) AND (CUDA_VERSION GREATER_EQUAL 11.6)))
endif()

foreach(TEST_OP ${TEST_OPS})
py_test_modules(${TEST_OP} MODULES ${TEST_OP})
py_test_modules(${TEST_OP} MODULES ${TEST_OP} ENVS "NVIDIA_TF32_OVERRIDE=0")
list(APPEND DIST_TEST_OPS ${TEST_OP})
set_tests_properties(${TEST_OP} PROPERTIES TIMEOUT 200)
set_tests_properties(${TEST_OP} PROPERTIES LABELS "RUN_TYPE=DIST")
Expand Down