# AWS says EKS, EFA and DeepEP lift MoE reinforcement-learning throughput 40%

> AWS's machine learning blog details an architecture combining Amazon EKS, Elastic Fabric Adapter and DeepEP to speed expert-parallel communication in large-scale MoE RL post-training, claiming 40% more throughput.

- **Topic**: Research
- **Published**: 2026-09-27T00:54:19.309Z
- **Canonical URL**: https://highsignal.sh/stories/aws-says-eks-efa-and-deepep-lift-moe-reinforcement-learning-throughput-40-8ccfdc0d

## Why It Matters

As MoE post-training shifts from compute-bound to communication-bound, infrastructure — not just algorithms — becomes the bottleneck for RLHF and GRPO pipelines.

## Key Findings & Analysis

### What AWS announced

- In an AWS Machine Learning Blog post, AWS describes an architecture for post-training Mixture-of-Experts (MoE) models with RLHF or GRPO at scale, combining Amazon Elastic Kubernetes Service (Amazon EKS) with Elastic Fabric Adapter (EFA) and DeepEP. AWS claims 40% more throughput from the setup, per the post's title and framing.
- AWS says the design targets three simultaneous problems: coordinating heterogeneous compute for rollout generation and policy training, sustaining high-throughput communication across hundreds of accelerators, and dynamically orchestrating subsystems to keep them in balance.
- AWS states the underlying constraint is that as MoE architectures grow sparser to cut inference costs, training becomes limited more by communication than by compute, with Expert Parallelism (EP) all-to-all token routing a key source of overhead on top of tensor, data and pipeline parallelism.
- AWS says the post covers how DeepEP optimizes expert-parallel communication over EFA, and frames the rollout-training loop as two concurrent workloads — throughput-oriented distributed inference plus tightly coupled policy training.
- Why it matters: AWS's claims are vendor-stated, not independently benchmarked in the supplied text, but they point to infrastructure configuration — networking and orchestration — as a lever on MoE RL efficiency alongside model and algorithm choices.

## Primary Sources & Citations

- [Scaling MoE reinforcement learning on Amazon EKS with EFA and DeepEP with 40% more throughput](https://aws.amazon.com/blogs/machine-learning/scaling-moe-reinforcement-learning-on-amazon-eks-with-efa-and-deepep-with-40-more-throughput/) — *AWS Machine Learning Blog* (Primary source)

---

[← Back to Headlines](https://highsignal.sh/) | [Daily Brief](https://highsignal.sh/brief) | [All stories](https://highsignal.sh/latest)