Gemmini: Enabling Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration

By Hasan Genç, Seah Kim, Alon Amid, Ameer Haj-Ali, Vighnesh Iyer, P. Prakash, Jerry Zhao, D. Grubb, Harrison Liew, H. Mao and 9 more. Design Automation Conference, 2019

367 citations recorded by the source corpus.

DNN accelerators are often developed and evaluated in isolation without considering the cross-stack, system-level effects in real-world environments. This makes it difficult to appreciate the impact of Systemon-Chip (SoC) resource contention, OS overheads, and programming-stack inefficiencies on overall performance/energy-efficiency. To address this challenge, we present Gemmini, an open-source, full-stack DNN accelerator generator. Gemmini generates a wide design-space of efficient ASIC accelerators from a flexible architectural template, together with flexible programming stacks and full SoCs with shared resources that capture system-level effects. Gemmini-generated accelerators have also been fabricated, delivering up to three orders-of-magnitude speedups over high-performance CPUs on various DNN benchmarks.

Subfields: Electrical and Electronic Engineering, Hardware and Architecture

View Paper: https://www.semanticscholar.org/paper/0e0a24ca2869e56ffe47a87d360b78b6994c8f58