Skip to content

removing the fuse distributed ops lowering pass for tegra platforms #3411

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

Merged
merged 2 commits into from
Mar 13, 2025
Merged

removing the fuse distributed ops lowering pass for tegra platforms #3411

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3c9b77f
removing the fuse distributed ops lowering pass for tegra platforms
apbose Feb 24, 2025
55535ca
utility function to detect tegra platform
apbose Feb 25, 2025
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
27 changes: 15 additions & 12 deletions py/torch_tensorrt/dynamo/lowering/passes/_aten_lowering_pass.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@

import torch
from torch_tensorrt.dynamo._settings import CompilationSettings
from torch_tensorrt.dynamo.utils import is_tegra_platform

from .accumulate_fp32_matmul import accumulate_fp32_matmul
from .constant_folding import constant_fold
Expand All @@ -15,18 +16,20 @@
from .repair_input_as_output import repair_input_as_output
from .replace_max_pool_with_indices import replace_max_pool_with_indices

ATEN_POST_LOWERING_PASSES = DynamoPassManager.build_from_passlist(
[
remove_input_alias_fixing_clones,
constant_fold,
repair_input_as_output,
fuse_prims_broadcast,
fuse_distributed_ops,
replace_max_pool_with_indices,
remove_assert_nodes,
accumulate_fp32_matmul,
]
)
pass_list = [
remove_input_alias_fixing_clones,
constant_fold,
repair_input_as_output,
fuse_prims_broadcast,
replace_max_pool_with_indices,
remove_assert_nodes,
accumulate_fp32_matmul,
]

if not is_tegra_platform():
pass_list.append(fuse_distributed_ops)

ATEN_POST_LOWERING_PASSES = DynamoPassManager.build_from_passlist(pass_list)

ATEN_PRE_LOWERING_PASSES = DynamoPassManager.build_from_passlist(
[
Expand Down
6 changes: 6 additions & 0 deletions py/torch_tensorrt/dynamo/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -806,3 +806,9 @@ def get_output_dtypes(output: Any, truncate_doulbe: bool = False) -> List[dtype]
f"got unexpected type {type(output)}, expected type is a torch.fx.node.Node or a tuple/list of torch.fx.node.Node"
)
return output_dtypes


def is_tegra_platform() -> bool:
if torch.cuda.get_device_capability() in [(8, 7), (7, 2)]:
return True
return False