Skip to content

Pull requests: deepseek-ai/DeepEP

New pull request New
56 Open 156 Closed
56 Open 156 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

shared-memory-management: auto-detect Fabric support
#559 opened Jan 8, 2026 by fengidri Loading…
Fix linking error with nvidia-nvshmem-cu12 3.5.19
#557 opened Jan 6, 2026 by windreamer Loading…
2
Add custom AG , remove the usage of some environments
#556 opened Jan 6, 2026 by Autumn1998 Loading…
release gil in intranode::dispatch
#555 opened Jan 4, 2026 by reyoung Loading…
Add architecture during device linking
#554 opened Jan 3, 2026 by rzambre Loading…
feat: add auto-detection for Blackwell GPU architecture (sm_100a)
#550 opened Dec 28, 2025 by yurekami Loading…
3 tasks
1
Improve .gitignore with common Python entries
#547 opened Dec 25, 2025 by yurekami Loading…
2 tasks done
Add missing return type hints to Buffer methods
#546 opened Dec 25, 2025 by yurekami Loading…
2 tasks
Fix grammar in installation docs
#545 opened Dec 25, 2025 by yurekami Loading…
Fix check_nvlink_connections called with None group
#543 opened Dec 25, 2025 by yurekami Loading…
2 tasks
Add printf buffer flush delay to EP_DEVICE_ASSERT
#542 opened Dec 25, 2025 by yurekami Loading…
2 tasks
Improve error messages for unsupported EP rank configurations
#541 opened Dec 25, 2025 by yurekami Loading…
Add __version__ and __all__ to package
#540 opened Dec 25, 2025 by yurekami Loading…
Add delay before trap() to allow printf buffer flush
#539 opened Dec 25, 2025 by yurekami Loading…
2 tasks
1
Use atomicMax instead of atomicAdd for wait_recv_cost_stats
#538 opened Dec 25, 2025 by yurekami Loading…
1
Add runtime warning for known num_tokens=512 bug
#537 opened Dec 25, 2025 by yurekami Loading…
2 tasks
Add NVSHMEM guards for low_latency mask buffer functions
#536 opened Dec 25, 2025 by yurekami Loading…
3 tasks
1
feat: add __version__ and __all__ exports to package
#535 opened Dec 25, 2025 by yurekami Loading…
2 of 3 tasks
Add DEEPEP_SKIP_KINETO env var to bypass kineto for nsys profiling
#534 opened Dec 25, 2025 by yurekami Loading…
2 tasks
Add barrier between cudaIpcCloseMemHandle and cudaFree in Buffer::destroy
#533 opened Dec 25, 2025 by yurekami Loading…
2 tasks
1
Fix TMA assertion for non-32-divisible hidden sizes in combine_token
#532 opened Dec 25, 2025 by yurekami Loading…
3 tasks
1
Fix EP128 nvl_buffer_size configuration in test_internode.py
#531 opened Dec 25, 2025 by yurekami Loading…
2 tasks
1
Fix the MultiStages for intranode combine and the offset_int4 for int…
#527 opened Dec 19, 2025 by cyhdmjzzy Loading…
Add py::module_local() to pybind11::class
#524 opened Dec 18, 2025 by ooooo-create Loading…
Fix combine correctness check in test_internode.py with correct eps
#523 opened Dec 18, 2025 by hexclover Loading…
ProTip! Exclude everything labeled bug with -label:bug.