Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / bytedance/flux issues and pull requests

#48 - [ENHANCEMENT] 你好，fp8有计划支持吗

Issue - State: open - Opened by zkyue 5 days ago - 2 comments
Labels: enhancement

#46 - [QUESTION] Not supported on A6000?

Issue - State: open - Opened by Zhuohao-Li 22 days ago - 3 comments

#45 - Update __init__.py

Pull Request - State: closed - Opened by zheng-ningxin about 1 month ago

#44 - [BUG] incorrect shape output from AGKernel.gather()

Issue - State: open - Opened by 152334H about 1 month ago - 1 comment

#43 - [QUESTION] Can flux run on RTX 4090?

Issue - State: open - Opened by qinghon about 1 month ago - 1 comment

#42 - [QUESTION]some questions about allgather+gemm

Issue - State: open - Opened by ChrisRanger about 1 month ago - 1 comment

#41 - Update __init__.py

Pull Request - State: closed - Opened by zheng-ningxin 2 months ago

#40 - [BUG] Illegal memory with multi-node

Issue - State: open - Opened by YJHMITWEB 2 months ago - 1 comment
Labels: bug

#40 - [BUG] Illegal memory with multi-node

Issue - State: open - Opened by YJHMITWEB 2 months ago - 1 comment
Labels: bug

#39 - [QUESTION] How does flux handle hardware resoureces competition?

Issue - State: open - Opened by chenhongyu2048 2 months ago - 2 comments
Labels: question

#39 - [QUESTION] How does flux handle hardware resoureces competition?

Issue - State: open - Opened by chenhongyu2048 2 months ago - 2 comments
Labels: question

#38 - [QUESTION] Can Gemm_V3 be used in SM80?

Issue - State: closed - Opened by ginowu 2 months ago - 6 comments
Labels: question

#35 - [ENHANCEMENT] support for gpu A40

Issue - State: closed - Opened by 1926627357 3 months ago - 3 comments
Labels: enhancement

#34 - [QUESTION] Why flux gemm_rs is not faster than torch?

Issue - State: open - Opened by hxdtest 3 months ago - 5 comments
Labels: question

#33 - [QUESTION] How to use nvshmem?

Issue - State: closed - Opened by chenhongyu2048 3 months ago - 8 comments
Labels: question

#32 - [BUG] `no_nvlink` branch failed to compile

Issue - State: closed - Opened by lucifer1004 3 months ago - 6 comments

#31 - [QUESTION]is there a plan to support int8?

Issue - State: closed - Opened by Rainlin007 3 months ago - 1 comment
Labels: question

#30 - [BUG] Failing to install byte-flux from pypi

Issue - State: open - Opened by tlrmchlsmth 3 months ago - 8 comments
Labels: bug

#29 - [QUESTION]How to run examples in pynvshmem

Issue - State: closed - Opened by TonyWu199 3 months ago - 1 comment

#28 - [BUG] Can't find nccl when building from source

Issue - State: open - Opened by KnowingNothing 4 months ago - 5 comments

#27 - [QUESTION] Are you planning on supporting FP8?

Issue - State: closed - Opened by MustafaFayez 4 months ago - 3 comments

#26 - Update README.md

Pull Request - State: closed - Opened by zheng-ningxin 4 months ago

#25 - Support pip installation

Pull Request - State: closed - Opened by zheng-ningxin 4 months ago

#24 - add torch version to the whl name

Pull Request - State: closed - Opened by zheng-ningxin 4 months ago

#23 - Support performance tunning for gemm-rs kernel on sm80

Pull Request - State: closed - Opened by zheng-ningxin 4 months ago

#22 - Remove pynvshmem import in gemm_rs_80.py

Pull Request - State: closed - Opened by tlrmchlsmth 4 months ago

#21 - Tune the AG performance for the llama-8b

Pull Request - State: closed - Opened by zheng-ningxin 4 months ago

#20 - Are there any difficulties in implementing gemm-allreduce?

Issue - State: open - Opened by Rainlin007 4 months ago - 2 comments
Labels: enhancement

#18 - zero out all the allocated shm buffer

Pull Request - State: closed - Opened by zheng-ningxin 4 months ago

#17 - [BUG] Incorrect results from flux.AGKernel for some problem shapes

Issue - State: closed - Opened by tlrmchlsmth 5 months ago - 15 comments
Labels: bug

#16 - Update README.md

Pull Request - State: closed - Opened by zheng-ningxin 5 months ago

#15 - Add more device types for the time estimation.

Pull Request - State: closed - Opened by zheng-ningxin 5 months ago

#14 - [BUG] Exception: not supported device NVIDIA H100 80GB HBM3

Issue - State: closed - Opened by wenscarl 5 months ago - 2 comments

#13 - using c10::intrusive_ptr<c10d::ProcessGroup> as argument from python

Pull Request - State: closed - Opened by houqi 5 months ago - 1 comment

#12 - fix the _allgather_base backend issue(issue11)

Pull Request - State: closed - Opened by zheng-ningxin 5 months ago

#11 - [BUG] RuntimeError: Could not retrieve or create the backend 2 for device type cuda

Issue - State: closed - Opened by tlrmchlsmth 5 months ago - 14 comments
Labels: bug

#10 - [BUG] Illegal memory access when fuse_reduction=False

Issue - State: closed - Opened by tlrmchlsmth 5 months ago - 5 comments
Labels: bug

#9 - Support IPC && SM90 version of AG-GEMM, GEMM-RS

Pull Request - State: closed - Opened by zheng-ningxin 5 months ago

#7 - [BUG] gemm and reduce-scatter are not overlapped

Issue - State: closed - Opened by wenscarl 5 months ago - 9 comments
Labels: question

#6 - Update BibTex

Pull Request - State: closed - Opened by wenlei-bao 5 months ago

#5 - Add arXiv paper link

Pull Request - State: closed - Opened by wenlei-bao 5 months ago

#4 - Reorganize and deduplicate files

Pull Request - State: closed - Opened by wenlei-bao 5 months ago

#3 - All gather and reduce scatter on SM80

Pull Request - State: closed - Opened by zheng-ningxin 5 months ago

#2 - add cutlass submodule and patches

Pull Request - State: closed - Opened by liwenchangbdbz 9 months ago

#1 - add issue template

Pull Request - State: closed - Opened by liwenchangbdbz 9 months ago