MoE-Inference-Bench: Performance Evaluation of Mixture of Expert Large Language and Vision Models

TL;DR

This paper systematically evaluates the performance of Mixture of Experts models for large language and vision models, analyzing hardware acceleration techniques and optimization strategies to improve inference efficiency.

Contribution

It provides a comprehensive benchmarking framework and analysis of various optimization techniques for MoE models on Nvidia H100 GPUs.

Findings

Performance varies significantly across configurations.

Optimization techniques like pruning and quantization improve throughput.

Insights guide efficient deployment of MoE models.

Abstract

Mixture of Experts (MoE) models have enabled the scaling of Large Language Models (LLMs) and Vision Language Models (VLMs) by achieving massive parameter counts while maintaining computational efficiency. However, MoEs introduce several inference-time challenges, including load imbalance across experts and the additional routing computational overhead. To address these challenges and fully harness the benefits of MoE, a systematic evaluation of hardware acceleration techniques is essential. We present MoE-Inference-Bench, a comprehensive study to evaluate MoE performance across diverse scenarios. We analyze the impact of batch size, sequence length, and critical MoE hyperparameters such as FFN dimensions and number of experts on throughput. We evaluate several optimization techniques on Nvidia H100 GPUs, including pruning, Fused MoE operations, speculative decoding, quantization, and various parallelization strategies. Our evaluation includes MoEs from the Mixtral, DeepSeek, OLMoE and Qwen families. The results reveal performance differences across configurations and provide insights for the efficient deployment of MoEs.