Towards Building Private LLMs: Exploring Multi-Node Expert Parallelism on Apple Silicon for Mixture-of-Experts Large Language Model
It addresses cost and scalability issues for personal or small group services, such as Apple Intelligence, but is incremental as it focuses on optimizing existing hardware and software for a specific use case.
This paper tackles the cost and scalability challenges of building private LLM systems by using a Mac Studio cluster with Apple M2 Ultra chips to host and accelerate a pretrained DBRX MoE model, achieving a 1.15 times cost efficiency improvement over state-of-the-art AI supercomputers with NVIDIA H100 GPUs.
Large Language Models (LLMs) have revolutionized Artificial Intelligence (AI) with significant advancements such as OpenAI's ChatGPT, Meta's Llama, and Databricks' DBRX. This paper addresses the cost and scalability challenges encountered when constructing private LLM systems for personal or small group services, as aimed by Apple Intelligence. A Mac Studio cluster with Apple's M2 Ultra chips is established as a cost-efficient solution to host and accelerate the pretrained DBRX model with the Mixture-of-Experts (MoE) architecture. Our performance analysis reveal that parallel execution of the model's experts across two to four machine nodes significantly reduces inference time. We find that computation time for the experts is comparable to the communication time for exchanging their outputs, emphasizing the importance of network latency over bandwidth. We also observe significant management overhead due to Apple software stack's memory management logic. Based on these findings, we develop optimization schemes to eliminate the memory management overhead. As a result, the Mac Studio cluster is 1.15 times more cost-efficient than the state-of-the-art AI supercomputer with NVIDIA H100 GPUs. In addition, we construct a performance model to estimate system performance under varying configurations, and the model provides valuable insights for designing private LLM systems.