Yin-Yang: Programming Abstractions for Cross-Domain Multi-Acceleration

Abstract

Field-programmable gate array (FPGA) accelerators offer performance and efficiency gains by narrowing the scope of acceleration to one algorithmic domain. However, real-life applications are often not limited to a single domain, which naturallymakes Cross-Domain Multi-Acceleration a crucial next step. The challenge is, existing FPGA accelerators are built upon their specific vertically specialized stacks, which prevents utilizing multiple accelerators from different domains. To that end, we propose a pair of dual abstractions, called Yin-Yang, which work in tandem and enable programmers to develop cross-domain applications using multiple accelerators on a FPGA. The Yin abstraction enables cross-domain algorithmic specification, while the Yang abstraction captures the accelerator capabilities. We also developed a dataflow virtual machine, dubbed Accelerator-Level Virtual Machine (XLVM), which transparently maps domain functions (Yin) to best-fit accelerator capabilities (Yang). With six real-world cross-domain applications, our evaluations showthat Yin-Yang unlocks 29.4x speedup, while the best single-domain acceleration achieves 12.0x.