Skip to content

Issues: bytedance/flux

Beta
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

17 Open 60 Closed
17 Open 60 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

[QUESTION]flux/src/moe_ag_scatter/ths_op/gemm_grouped_v3_ag_scatter.cc里的写stream很耗时，有优化方法吗？CU_CHECK(CUStreamWriteValue(this->cp_stream_intra_node,(CUdeviceptr)(ptr_offset(barrier_block.get(), src_rank * sizeof(int))),1,CU_STREAM_WRITE_VALUE_DEFAULT));
#113 opened Apr 29, 2025 by jinchen89
[QUESTION]flux/src/moe_ag_scatter/ths_op/gemm_grouped_v3_ag_scatter.cc你好，如果我需要在moe_ag的allgather之后对input_buffer做一个transpose(0,1),需要怎么修改这里的代码？需要改flux/src/moe_ag_scatter/sm90_gemm_array_warpspecialized_cooperative.hpp里面的代码吗？
#112 opened Apr 28, 2025 by jinchen89
@ZSL98
[QUESTION]你好，flux/test/python/moe_ag_scatter/test_moe_ag.py这个需要在8gpu的时候设置tp size=2需要怎么改？
#110 opened Apr 19, 2025 by jinchen89
@ZSL98
3
[QUESTION]您好，请教一下如果我需要在moe_ag的allgather通信之后对ctx.inputs增加一个transpose(0,1)，需要修改哪处代码？
#103 opened Apr 15, 2025 by jinchen89
1
[QUESTION] Gemm +RS on 8xH100
#100 opened Apr 12, 2025 by rajagond
@wenlei-bao
1
[QUESTION]你好，我在nv H20上跑./launch.sh test/python/gemm_rs/test_gemm_rs.py 16384 4096 16384 --dtype=bfloat16 --iters=5报了这个错，H800上同样的镜像和命令没有问题？是H20有啥不一样吗？
#89 opened Apr 6, 2025 by jinchen89
@ZSL98
6
[BUG] Fuse Reduction on SM 90 bug Something isn't working
#86 opened Apr 5, 2025 by rajagond
@zheng-ningxin
1
[QUESTION] E2E Overlap: Flux design
#80 opened Mar 30, 2025 by rajagond
@wenlei-bao
8
[QUESTION]flux的性能比torch略慢一些？
#79 opened Mar 28, 2025 by jinchen89
@kongroo
2
[BUG] Fp8 Runtime Error: "bad any_cast"
#76 opened Mar 26, 2025 by DXHPC
7
[QUESTION]IF FLUX supports RoCE NIC?
#51 opened Feb 15, 2025 by xuzhenguoloveyjh
1
[QUESTION] Not supported on A6000?
#46 opened Oct 26, 2024 by Zhuohao-Li
@zheng-ningxin
6
[BUG] Illegal memory with multi-node bug Something isn't working
#40 opened Sep 8, 2024 by YJHMITWEB
@wenlei-bao
4
[QUESTION] The gemm time on GPU of different rank under tp8 is very different , and cause low performance question Further information is requested
#36 opened Aug 22, 2024 by Rainlin007
@zheng-ningxin
8
[QUESTION] Why flux gemm_rs is not faster than torch? question Further information is requested
#34 opened Aug 21, 2024 by hxdtest
@zheng-ningxin
6
[BUG] Failing to install byte-flux from pypi bug Something isn't working
#30 opened Aug 13, 2024 by tlrmchlsmth
@zheng-ningxin
12
Are there any difficulties in implementing gemm-allreduce? enhancement New feature or request
#20 opened Jul 15, 2024 by Rainlin007
@zheng-ningxin
2
ProTip! Adding no:label will show everything without a label.