跳到正文
RCreddit.com·
暂不在当前实时榜单

Finally got Qwen 3.8 Next running on my v100 6gpu setup (TP2 PP3)

AI 摘要

A user successfully deployed Qwen 3.8 Next on a v100 6GPU setup, configured with TP2 PP3. Performance metrics show varying token generation speeds across different input lengths, ranging from 1,389 tok/s for 1K tokens to 2,759 tok/s for 131K tokens. The system achieved a maximum speed of 4,679 tok/s at 16,384 (16K) input length. The user expressed satisfaction with the system's stability, thermal management, and low noise levels.

为什么是这条

This report is the first to detail Qwen 3.8 Next's performance on a v100 6GPU setup, providing specific token generation speeds across various input lengths, unlike general benchmarks.

时间与来源

时间显示为 UTC

显示时区:UTC

本地时区尚不可用,暂时显示 UTC。

收录当时偏移:UTC+02026年9月20日 11:01 UTC

收录
2026年9月20日 11:01
来源类型
开发者社区

本站未收录正文。

前往源站阅读 →
来源·reddit.com