BENCHIP: Benchmarking Intelligence Processors

TL;DR

BENCHIP introduces a comprehensive benchmark suite and methodology for fair and representative evaluation of diverse intelligence processors, addressing current limitations in existing benchmarks.

Contribution

It provides a standardized benchmarking framework with micro- and macrobenchmarks and evaluation metrics for intelligence hardware comparison.

Findings

BENCHIP effectively differentiates hardware performance.

It enables fair comparison across CPUs, GPUs, and accelerators.

The benchmark suite is suitable for system optimization and bottleneck analysis.

Abstract

The increasing attention on deep learning has tremendously spurred the design of intelligence processing hardware. The variety of emerging intelligence processors requires standard benchmarks for fair comparison and system optimization (in both software and hardware). However, existing benchmarks are unsuitable for benchmarking intelligence processors due to their non-diversity and nonrepresentativeness. Also, the lack of a standard benchmarking methodology further exacerbates this problem. In this paper, we propose BENCHIP, a benchmark suite and benchmarking methodology for intelligence processors. The benchmark suite in BENCHIP consists of two sets of benchmarks: microbenchmarks and macrobenchmarks. The microbenchmarks consist of single-layer networks. They are mainly designed for bottleneck analysis and system optimization. The macrobenchmarks contain state-of-the-art industrial networks, so as to offer a realistic comparison of different platforms. We also propose a standard benchmarking methodology built upon an industrial software stack and evaluation metrics that comprehensively reflect the various characteristics of the evaluated intelligence processors. BENCHIP is utilized for evaluating various hardware platforms, including CPUs, GPUs, and accelerators. BENCHIP will be open-sourced soon.