Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Park, Jihoon, Choe, Jeongin, Kim, Dohyun, Kim, Jae-Joon
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2501.06780
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910781863362560
author Park, Jihoon
Choe, Jeongin
Kim, Dohyun
Kim, Jae-Joon
author_facet Park, Jihoon
Choe, Jeongin
Kim, Dohyun
Kim, Jae-Joon
contents Recently, crossbar array based in-memory accelerators have been gaining interest due to their high throughput and energy efficiency. While software and compiler support for the in-memory accelerators has also been introduced, they are currently limited to the case where all weights are assumed to be on-chip. This limitation becomes apparent with the significantly increasing network sizes compared to the in-memory footprint. Weight replacement schemes are essential to address this issue. We propose COMPASS, a compiler framework for resource-constrained crossbar-based processing-in-memory (PIM) deep neural network (DNN) accelerators. COMPASS is specially targeted for networks that exceed the capacity of PIM crossbar arrays, necessitating access to external memories. We propose an algorithm to determine the optimal partitioning that divides the layers so that each partition can be accelerated on chip. Our scheme takes into account the data dependence between layers, core utilization, and the number of write instructions to minimize latency, memory accesses, and improve energy efficiency. Simulation results demonstrate that COMPASS can accommodate much more networks using a minimal memory footprint, while improving throughput by 1.78X and providing 1.28X savings in energy-delay product (EDP) over baseline partitioning methods.
format Preprint
id arxiv_https___arxiv_org_abs_2501_06780
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle COMPASS: A Compiler Framework for Resource-Constrained Crossbar-Array Based In-Memory Deep Learning Accelerators
Park, Jihoon
Choe, Jeongin
Kim, Dohyun
Kim, Jae-Joon
Hardware Architecture
Distributed, Parallel, and Cluster Computing
Emerging Technologies
Machine Learning
Programming Languages
Recently, crossbar array based in-memory accelerators have been gaining interest due to their high throughput and energy efficiency. While software and compiler support for the in-memory accelerators has also been introduced, they are currently limited to the case where all weights are assumed to be on-chip. This limitation becomes apparent with the significantly increasing network sizes compared to the in-memory footprint. Weight replacement schemes are essential to address this issue. We propose COMPASS, a compiler framework for resource-constrained crossbar-based processing-in-memory (PIM) deep neural network (DNN) accelerators. COMPASS is specially targeted for networks that exceed the capacity of PIM crossbar arrays, necessitating access to external memories. We propose an algorithm to determine the optimal partitioning that divides the layers so that each partition can be accelerated on chip. Our scheme takes into account the data dependence between layers, core utilization, and the number of write instructions to minimize latency, memory accesses, and improve energy efficiency. Simulation results demonstrate that COMPASS can accommodate much more networks using a minimal memory footprint, while improving throughput by 1.78X and providing 1.28X savings in energy-delay product (EDP) over baseline partitioning methods.
title COMPASS: A Compiler Framework for Resource-Constrained Crossbar-Array Based In-Memory Deep Learning Accelerators
topic Hardware Architecture
Distributed, Parallel, and Cluster Computing
Emerging Technologies
Machine Learning
Programming Languages
url https://arxiv.org/abs/2501.06780