Koios: A Deep Learning Benchmark Suite for FPGA Architecture and CAD Research

TL;DR

Koios introduces a comprehensive FPGA benchmark suite tailored for deep learning workloads, enabling more realistic and challenging evaluations of FPGA architecture and CAD tools to optimize deep learning performance.

Contribution

This work presents Koios, a new FPGA benchmark suite with 19 diverse DL circuits, addressing the inadequacies of existing benchmarks for DL-specific FPGA research.

Findings

Benchmarks are larger, more data parallel, and utilize more FPGA features.

Koios benchmarks have higher DSP, BRAM densities, and netlist primitives.

Case studies demonstrate improved architectural exploration for DL-optimized FPGAs.

Abstract

With the prevalence of deep learning (DL) in many applications, researchers are investigating different ways of optimizing FPGA architecture and CAD to achieve better quality-of-results (QoR) on DL-based workloads. In this optimization process, benchmark circuits are an essential component; the QoR achieved on a set of benchmarks is the main driver for architecture and CAD design choices. However, current academic benchmark suites are inadequate, as they do not capture any designs from the DL domain. This work presents a new suite of DL acceleration benchmark circuits for FPGA architecture and CAD research, called Koios. This suite of 19 circuits covers a wide variety of accelerated neural networks, design sizes, implementation styles, abstraction levels, and numerical precisions. These designs are larger, more data parallel, more heterogeneous, more deeply pipelined, and utilize more FPGA architectural features compared to existing open-source benchmarks. This enables researchers to pin-point architectural inefficiencies for this class of workloads and optimize CAD tools on more realistic benchmarks that stress the CAD algorithms in different ways. In this paper, we describe the designs in our benchmark suite, present results of running them through the Verilog-to-Routing (VTR) flow using a recent FPGA architecture model, and identify key insights from the resulting metrics. On average, our benchmarks have 3.7x more netlist primitives, 1.8x and 4.7x higher DSP and BRAM densities, and 1.7x higher frequency with 1.9x more near-critical paths compared to the widely-used VTR suite. Finally, we present two example case studies showing how architectural exploration for DL-optimized FPGAs can be performed using our new benchmark suite.