PIM-DL: Expanding the Applicability of Commodity DRAM-PIMs for Deep Learning via Algorithm-System Co-Optimization
Cong Li, Zhe Zhou, Yang Wang, Fan Yang, Ting Cao, Mao Yang, Yun Liang, Guangyu Sun
Abstract
DRAM-based processing-in-memory (DRAM-PIM) has gained commercial prominence in recent years. However, their integration for deep learning acceleration poses inherent challenges. Existing DRAM-PIMs are limited in computational capabilities, primarily applicable for element-wise and GEMV operators. Unfortunately, these operators contribute only a small portion of the execution time in most DNN workloads. Current systems still necessitate powerful hosts to handle a significant portion of compute-heavy operators.
Topics & Concepts
DramComputer scienceDeep learningParallel computingCommodityEmbedded systemComputer engineeringArtificial intelligenceComputer architectureComputer hardwareMarket economyEconomicsAdvanced Memory and Neural ComputingFerroelectric and Negative Capacitance DevicesAdvanced Neural Network Applications