2022/03/22 by Ahmet Caner Yüzügüler, Canberk Sönmez, Yüzügüler, Ahmet Caner +9
Biochemistry, Genetics and Molecular Biology · Computer Science · Engineering · #Advanced Memory and Neural Computing #Advanced biosensing and bioanalysis techniques #FOS: Computer and information sciences #Hardware Architecture (cs.AR) #Interconnection Networks and Systems #Machine Learning (cs.LG)
paper · pdf · doi:10.48550/arxiv.2203.11540
openalex publication_date 2022/03/22 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Multi-pod systolic arrays are emerging as the architecture of choice in DNN inference accelerators. Despite their potential, designing multi-pod systolic arrays to maximize effective throughput/Watt (i.e., throughput/Watt adjusted when accounting for array utilization) poses a unique set of challenges. In this work, we study three key pillars in multi-pod systolic array designs, namely array granularity, interconnect, and tiling. We identify optimal array granularity across workloads and show that state-of-the-art commercial accelerators use suboptimal array sizes for single-tenancy workloads. We, then evaluate the bandwidth/latency trade-offs in interconnects and show that Butterfly networks offer a scalable topology for accelerators with a large number of pods. Finally, we introduce a novel data tiling scheme with custom partition size to maximize utilization in optimally sized pods. We propose Scale-out Systolic Arrays, a multi-pod inference accelerator for both single- and multi-tenancy based on these three pillars. We show that SOSA exhibits scaling of up to 600 TeraOps/s in effective throughput for state-of-the-art DNN inference workloads, and outperforms state-of-the-art multi-pod accelerators by a factor of 1.5x.