Accelerating Neural Networks for Large Language Models and Graph Processing with Silicon Photonics
This addresses the problem of inefficient hardware acceleration for complex AI models like LLMs and graph processing, offering significant performance gains for researchers and practitioners in NLP, computer vision, and graph applications, though it is incremental as it builds on existing hardware acceleration methods with a new technology.
The paper tackled the challenge of accelerating transformer neural networks for large language models and graph neural networks for graph processing by developing novel hardware accelerators based on silicon photonics, achieving at least 10.2x throughput improvement and 3.8x better energy efficiency over state-of-the-art electronic accelerators.
In the rapidly evolving landscape of artificial intelligence, large language models (LLMs) and graph processing have emerged as transformative technologies for natural language processing (NLP), computer vision, and graph-structured data applications. However, the complex structures of these models pose challenges for acceleration on conventional electronic platforms. In this paper, we describe novel hardware accelerators based on silicon photonics to accelerate transformer neural networks that are used in LLMs and graph neural networks for graph data processing. Our analysis demonstrates that both hardware accelerators achieve at least 10.2x throughput improvement and 3.8x better energy efficiency over multiple state-of-the-art electronic hardware accelerators designed for LLMs and graph processing.