Skip to content

[PyTorch] Expert Parallelism: PyTorch wrapper + autograd ops with symm-mem zero-copy #3035

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

Draft
phu0ngng wants to merge 5 commits into
base: main
Choose a base branch
from
Draft

[PyTorch] Expert Parallelism: PyTorch wrapper + autograd ops with symm-mem zero-copy #3035

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
44a8a49
Expert Parallelism: common C API + NCCL EP backend
phu0ngng May 21, 2026
63afb52
Bump NCCL to v0.1
phu0ngng May 22, 2026
099857f
ep: revert cudnn-frontend bump, use NCCL EP v0.1 magic-aware INIT mac…
phu0ngng May 22, 2026
b2931a6
Expert Parallelism: common C API, NCCL EP backend, PyTorch autograd o…
phu0ngng May 22, 2026
f19ceff
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] May 22, 2026
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
6 changes: 5 additions & 1 deletion .gitmodules
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,4 +6,8 @@
url = https://github.com/NVIDIA/cudnn-frontend.git
[submodule "3rdparty/cutlass"]
path = 3rdparty/cutlass
url = https://github.com/NVIDIA/cutlass.git
url = https://github.com/NVIDIA/cutlass.git
[submodule "3rdparty/nccl"]
path = 3rdparty/nccl
url = https://github.com/NVIDIA/nccl.git
branch = v2.30u1
1 change: 1 addition & 0 deletions 3rdparty/nccl
Open in desktop
Submodule nccl added at 6a9bc9
6 changes: 6 additions & 0 deletions build_tools/pytorch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -76,6 +76,12 @@ def setup_pytorch_extension(

setup_mpi_flags(include_dirs, cxx_flags)

# Mirror the NCCL EP gate from setup.py / common CMake. When disabled, the
# ep.cpp source no-ops at the #ifdef boundary; without the define it would
# produce undefined references to nvte_ep_*.
if bool(int(os.getenv("NVTE_BUILD_WITH_NCCL_EP", "1"))):
cxx_flags.append("-DNVTE_WITH_NCCL_EP")

library_dirs = []
libraries = []
if bool(int(os.getenv("NVTE_ENABLE_NVSHMEM", 0))):
Expand Down
Loading
Loading