expert-parallelism

Tag

Cards List
#expert-parallelism

Director: Accelerating Distributed MoE Serving via Online Proactive Expert Placement

arXiv cs.LG · yesterday Cached

This paper introduces Director, a distributed MoE serving system that minimizes end-to-end latency using prediction-driven, online proactive expert placement. It employs a lightweight predictor and a relaxation-based optimizer to achieve up to 55% latency reduction for models like Mistral, DeepSeek, and Qwen.

0 favorites 0 likes
#expert-parallelism

@h100envy: Ex-Berkeley PhD who leads SGLang at xAI explained how they serve Grok on 100K GPUs in 23 minutes - better than $2000 in…

X AI KOLs Timeline · 2026-07-06 Cached

A former Berkeley PhD who leads SGLang at xAI explains how they serve Grok on 100K GPUs using split prefill/decode, expert sharding, and communication/computation overlap to achieve DeepSeek-API-killing prices.

0 favorites 0 likes
#expert-parallelism

MACS: Modality-Aware Capacity Scaling for Efficient Multimodal MoE Inference

arXiv cs.LG · 2026-05-08 Cached

MACS is a training-free inference framework that mitigates the straggler effect in expert parallelism for multimodal MoE MLLMs by introducing entropy-weighted load and dynamic modality-adaptive capacity mechanisms.

0 favorites 0 likes
#expert-parallelism

Federation of Experts: Communication Efficient Distributed Inference for Large Language Models

Hugging Face Daily Papers · 2026-05-07 Cached

Federation of Experts (FoE) restructures mixture-of-experts blocks into clusters that process KV heads independently, eliminating inter-node communication bottlenecks and improving inference throughput and latency by up to 5.2x while maintaining generation quality.

0 favorites 0 likes
← Back to home

Submit Feedback