Pushing up to the Limit of Memory Bandwidth and Capacity Utilization for Efficient LLM Decoding on Embedded FPGA

Abstract: The extremely high computational and storage demands of LLMs have excluded most edge devices, which were widely used for efficient machine learning, from being viable options. A typical edge device usually only has 4GB of memory capacity and a bandwidth of less than 20GB/s, while a LLM quantized to 4-bit precision with 7B parameters already requires 3.5GB of capacity, and its decoding process is purely bandwidth-bound. In this paper, we aim to explore these limits by proposing a hardware accelerator for LLM inference on the Zynq-based KV260 platform, equipped with 4GB of 64-bit 2400Mbps DDR4 memory. We successfully deploy a LLaMA2-7B model, achieving a decoding speed of around 5 token/s, utilizing 93.3% of the memory capacity and reaching 85% decoding speed of the theoretical memory bandwidth limit. To fully reserve the memory capacity for model weights and key-value cache, we develop the system in a bare-metal environment without an operating system. To fully reserve the bandwidth for model weight transfers, we implement a customized dataflow with an operator fusion pipeline and propose a data arrangement format that can maximize the data transaction efficiency. This research marks the first attempt to deploy a 7B level LLM on a standalone embedded field programmable gate array (FPGA) device. It provides key insights into efficient LLM inference on embedded FPGA devices and provides guidelines for future architecture design.