Skip to content

Issues: NVIDIA/nccl

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

708 Open 711 Closed
708 Open 711 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
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

directSend in treeUpDown need update
#1628 opened Mar 3, 2025 by visualxu
unsafe shmem syncthreads
#1627 opened Mar 3, 2025 by visualxu
allreduce issues duplicated irecv under network plugin
#1626 opened Mar 3, 2025 by YangZhou1997
Potential memory leak in net_ib.cc
#1625 opened Mar 3, 2025 by GeofferyGeng
Does the new profiler api of profiler plugin save the dump file when app is completed?
#1624 opened Mar 2, 2025 by Xiaoaier-Z-L
3
Question re: profiler plugin example results for AllReduce
#1623 opened Feb 28, 2025 by ilpolyak
1
Error when debugging nccl source code using cuda-gdb.
#1622 opened Feb 28, 2025 by Linus-Voss
[bug]it look like a potential bug that use a local variable for asynchronous copy
#1621 opened Feb 28, 2025 by shizhixing
1
Question about network plugin API : is there a finishing call ?
#1620 opened Feb 27, 2025 by EmmanuelBRELLE
1
connection timeout when cross machine
#1619 opened Feb 27, 2025 by Suki001
3
Add field in config to enable/disable CollNet
#1616 opened Feb 25, 2025 by lw
Do we have plans to support IBGDA technology ？
#1615 opened Feb 25, 2025 by visualxu
1
4090's four card has high P2P bandwidth(52GB/s), but the bandwidth of the lstopo intermediate node is low(8GB/s)
#1614 opened Feb 24, 2025 by Linus-Voss
[bug] bootstrap may fail when connection close before sending 8 bytes
#1612 opened Feb 22, 2025 by fishautumn
1
1
The two-threaded all-to-all communication timeout error
#1610 opened Feb 18, 2025 by wczhao
3
NCCL timeout
#1609 opened Feb 17, 2025 by SabrinaZhuangxx
1
2.21.5 hangs, after using P2P disable, SHM disable, memory allocations. Reproducing Error 2 unhandled and Error 3 queue developers
#1608 opened Feb 16, 2025 by HyouinSchoolAcc
1
How to trigger several independent communications（e.g allgather) simultaneously?
#1607 opened Feb 16, 2025 by Ind1x1
5
NCCL fails to run with Tailscale over NAT, works in LAN
#1606 opened Feb 16, 2025 by helloburke
1
Segfault when using ncclCommSplit with multiple threads and non-blocking init
#1605 opened Feb 14, 2025 by lw
8
4090 with P2P, alltoall is too low
#1603 opened Feb 12, 2025 by artetaout
3
[Question] On data path of alltoall/alltoallv
#1601 opened Feb 11, 2025 by GHGmc2
2
Why is only part of the NICs used during allreduce, and why are inter-node connections not on the same rail even when I explicitly specify using all NICs?
#1600 opened Feb 11, 2025 by FortPercent
2
Potential bug in group.cc
#1599 opened Feb 8, 2025 by songhexiang
2
why NCCL_MAX_NCHANNELS cannot limit ncclDevKernel_SendRecv grid size
#1598 opened Feb 8, 2025 by Graham1025
7
ProTip! Type g i on any issue or pull request to go back to the issue listing page.