跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+479新增论文
+361其中 AI 相关
+64已有论文修订
+95新增 AI 导读

arXiv 最近完成扫描:2026-11-16 01:34 · 全站最近入库:2026-11-16 21:56各信源进度 →

统计口径与数据新鲜度

统计 2026-11-16 03:09 — 2026-11-17 03:09(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-16 01:34最近发现新增版本:2026-11-16 21:56库内论文最新更新时间:2026-11-14 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

23,827 已收录论文19,941 AI 相关6,922 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 1 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-16 01:33(北京时间)
新论文 · v1arXiv:2611.11402

A Per-Step Load-Balancing Engine for Expert-Parallel MoE Serving

Jingyuan Li, Maolin Wang, Hayden So

原文摘要

In expert-parallel Mixture-of-Experts serving, a layer finishes only when the most loaded GPU of its group does. Token routing shifts at every decode step and concentrates on a few experts, so most of the group's reserved GPU time waits on stragglers. Recovering that time means re-deciding, at every layer of every decode step, how each replicated expert's tokens divide between its copies. The exact linear-program solvers that make this decision optimally take longer than many of the layers they serve, and their working set outgrows GPU shared memory at deployment scale. We introduce Relief, a bidirectional pairwise relaxation that makes the linear program unnecessary, coming within a fraction of a point of the per-step optimum in a few passes of integer halving over the token counts. We build and measure three deployments, CUDA kernels on the serving GPU, the same verified engine on an FPGA over PCIe, and a fixed-function block beside the accelerator. On decode routing from three open MoE families, Relief holds the balance near the per-step optimum at every group size and layer time measured. Scored as effective utilization relative to that optimum, the kernel alone matches or exceeds every deployable scheme at every group size, from the hundred-microsecond layer on. The on-die engine widens that lead at every point, recovering up to 37.3 percentage points over the deployed static placement, while adding 0.02% of a current accelerator's transistors.

cs.AR · cs.DC