Pith. sign in

REVIEW 19 cited by

MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2504.02263 v4 pith:HPCYHLSY submitted 2025-04-03 cs.DC cs.LG

MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism

classification cs.DC cs.LG
keywords megascale-inferparallelismattentionmodulesoverheadcommunicationdatadisaggregated
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Mixture-of-Experts (MoE) showcases tremendous potential to scale large language models (LLMs) with enhanced performance and reduced computational complexity. However, its sparsely activated architecture shifts feed-forward networks (FFNs) from being compute-intensive to memory-intensive during inference, leading to substantially lower GPU utilization and increased operational costs. We present MegaScale-Infer, an efficient and cost-effective system for serving large-scale MoE models. MegaScale-Infer disaggregates attention and FFN modules within each model layer, enabling independent scaling, tailored parallelism strategies, and heterogeneous deployment for both modules. To fully exploit disaggregation in the presence of MoE's sparsity, MegaScale-Infer introduces ping-pong pipeline parallelism, which partitions a request batch into micro-batches and shuttles them between attention and FFNs for inference. Combined with distinct model parallelism for each module, MegaScale-Infer effectively hides communication overhead and maximizes GPU utilization. To adapt to disaggregated attention and FFN modules and minimize data transmission overhead (e.g., token dispatch), MegaScale-Infer provides a high-performance M2N communication library that eliminates unnecessary GPU-to-CPU data copies, group initialization overhead, and GPU synchronization. Experimental results indicate that MegaScale-Infer achieves up to 1.90x higher per-GPU throughput than state-of-the-art solutions.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 19 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Sieve: Dynamic Expert-Aware PIM Acceleration for Evolving Mixture-of-Experts Models

    cs.AR 2026-05 conditional novelty 8.0

    Sieve dynamically schedules MoE experts across GPU and PIM hardware to handle bimodal token distributions, achieving 1.3x to 1.6x gains in throughput and interactivity over static prior PIM systems on three large models.

  2. SmoothAgent: Efficient Long-Horizon LLM-Based Agent Serving with Lookahead Context Engineering

    cs.DC 2026-06 unverdicted novelty 7.0

    SmoothAgent introduces lookahead context engineering to eliminate transformation overhead in LLM agents, reducing TTFT by up to 11.9x through proactive KV cache preparation.

  3. ViBE: Co-Optimizing Workload Skew and Hardware Variability for MoE Serving

    cs.DC 2026-05 unverdicted novelty 7.0

    ViBE co-optimizes expert placement with measured GPU performance variability in MoE inference to cut execution-time imbalance, delivering 14% better SLO attainment and up to 45% lower P90 TTFT.

  4. Frontier: Towards Comprehensive and Accurate LLM Inference Simulation

    cs.DC 2026-05 unverdicted novelty 7.0

    Frontier is a new discrete-event simulator for disaggregated LLM serving that incorporates co-location, PDD, AFD, and optimizations, achieving under 4% throughput error and large reductions in latency prediction error...

  5. Surviving Partial Rank Failures in Wide Expert-Parallel MoE Inference

    cs.DC 2026-05 unverdicted novelty 7.0

    EEP makes wide expert-parallel MoE serving survive single-rank failures with an 11s recovery pause, 8s reintegration pause, and throughput restored to 95% of pre-fault level within 52s while staying within 4.4% of a f...

  6. Analytical Provisioning for Attention-FFN Disaggregated LLM Serving under Stochastic Workloads

    cs.LG 2026-01 unverdicted novelty 7.0

    A renewal-reward analysis yields a closed-form mean-field rule for the optimal Attention/FFN provisioning ratio in disaggregated LLM serving that accounts for stochastic KV-cache growth and matches simulation optima w...

  7. UBEP: Re-architecting Expert Parallelism Communication Library for Production Superpods

    cs.DC 2026-07 conditional novelty 6.0

    UBEP replaces BSP All-to-All for MoE on multi-tier superpods with dependency-driven kernel decomposition, topology-aware token scheduling, and Data-as-Flag atomics, cutting All-to-All latency up to 52.4% and TPOT up to 11.1%.

  8. UBEP: Re-architecting Expert Parallelism Communication Library for Production Superpods

    cs.DC 2026-07 accept novelty 6.0

    UBEP re-architects MoE All-to-All communication for superpods via kernel decomposition, topology-aware scheduling, and data-as-flag synchronization, achieving up to 52.4% latency reduction on CM384 hardware.

  9. Think Before You Grid-Search: Floor-First Triage for LLM Serving

    cs.PF 2026-07 conditional novelty 6.0

    A five-dimensional resource-vector floor model computes latency bounds and capacity walls for LLM serving, predicting when TP16 or EP16+DP attention layouts dominate based on operating point.

  10. Think Before You Grid-Search: Floor-First Triage for LLM Serving

    cs.PF 2026-07 conditional novelty 6.0

    LLM serving should triage by five-resource analytical floors and wall ordering, not grid search; on 16×H20, TP16 is capacity-capped at ~70 while EP+DP attention reaches ~644 concurrent 8K requests.

  11. DisagMoE: Computation-Communication overlapped MoE Training via Disaggregated AF-Pipe Parallelism

    cs.LG 2026-05 unverdicted novelty 6.0

    DisagMoE achieves up to 1.8x faster MoE training by disaggregating attention and FFN layers into disjoint GPU groups with a multi-stage uni-directional pipeline and roofline-based bandwidth balancing.

  12. AMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving

    cs.AR 2026-04 unverdicted novelty 6.0

    AMMA is a memory-centric multi-chiplet architecture using HBM-PNM cubes, custom logic dies, hybrid parallelism, and reordered collectives that delivers 15.5X lower attention latency and 6.9X lower energy than NVIDIA H...

  13. CascadeInfer: Length-Aware Scheduling of LLM Serving with Low Latency and Load Balancing

    cs.DC 2025-12 conditional novelty 6.0

    CascadeInfer partitions LLM instances into length-specialized groups, uses dynamic programming for stage partitioning, and applies runtime refinement plus decentralized load balancing to cut latency and raise throughput.

  14. Patterns behind Chaos: Forecasting Data Movement for Efficient Large-Scale MoE LLM Inference

    cs.DC 2025-10 conditional novelty 6.0

    Comprehensive profiling of expert selection in frontier MoE models reveals temporal and spatial patterns that enable 6.6x speedup on wafer-scale GPUs and 1.25x on existing systems via targeted optimizations.

  15. KernelFlume: Elastic Core-Attention Scaling for Agentic Long-Context Decoding

    cs.DC 2026-06 unverdicted novelty 5.0

    KernelFlume presents a disaggregated decode architecture that separates core attention from projection/FFN paths to enable elastic scaling of attention nodes, reporting up to 61% lower cost per million tokens versus f...

  16. How Far Can Disaggregation Go? A Design-Space Exploration of Attention-FFN Disaggregation for Efficient MoE LLM Serving

    cs.LG 2026-05 unverdicted novelty 5.0

    Operator-level attention-FFN disaggregation enables ~4k tokens/s throughput for DeepSeek-V3.2 under tight TTFT/TPOT SLOs where chunked-prefill and prefill-decode baselines cannot.

  17. UniEP: Unified Expert-Parallel MoE MegaKernel for LLM Training

    cs.DC 2026-04 unverdicted novelty 5.0

    UniEP fuses MoE communication and computation into unified MegaKernels with deterministic token ordering, delivering 1.03x-1.38x speedups over prior work while preserving training accuracy.

  18. Understanding and Improving Communication Performance in Multi-node LLM Inference

    cs.DC 2025-11 conditional novelty 5.0

    Performance analysis of multi-node LLM inference identifies all-reduce bottlenecks and introduces NVRAR hierarchical all-reduce achieving 1.9-3.6x lower latency than NCCL and up to 1.72x end-to-end batch latency reduc...

  19. Position: LLM Serving Needs Mathematical Optimization and Algorithmic Foundations, Not Just Heuristics

    cs.DC 2026-05 accept novelty 4.0

    LLM serving requires mathematical optimization and algorithms with provable guarantees rather than generic heuristics that fail unpredictably on LLM workloads.