High Performance Scalable FPGA Accelerator for Deep Neural Networks
This work addresses the need for efficient and versatile hardware acceleration for AI inference, offering a competitive alternative to existing platforms.
The paper tackles the problem of accelerating deep neural network inference by proposing an FPGA-based accelerator using INT-8-2 compute, achieving 5 AI-TOPS on Arria10 and projecting 76 AI-TOPS on Stratix10, which exceeds CPU and GPU performance and approaches ASIC levels.
Low-precision is the first order knob for achieving higher Artificial Intelligence Operations (AI-TOPS). However the algorithmic space for sub-8-bit precision compute is diverse, with disruptive changes happening frequently, making FPGAs a natural choice for Deep Neural Network inference, In this work we present an FPGA-based accelerator for CNN inference acceleration. We use {\it INT-8-2} compute (with {\it 8 bit} activation and {2 bit} weights) which is recently showing promise in the literature, and which no known ASIC, CPU or GPU natively supports today. Using a novel Adaptive Logic Module (ALM) based design, as a departure from traditional DSP based designs, we are able to achieve high performance measurement of 5 AI-TOPS for {\it Arria10} and project a performance of 76 AI-TOPS at 0.7 TOPS/W for {\it Stratix10}. This exceeds known CPU, GPU performance and comes close to best known ASIC (TPU) numbers, while retaining the versatility of the FPGA platform for other applications.