HePGA: A Heterogeneous Processing-in-Memory based GNN Training Accelerator

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ogbogu, Chukwufumnanya, Narang, Gaurav, Joardar, Biresh Kumar, Doppa, Janardhan Rao, Chakrabarty, Krishnendu, Pande, Partha Pratim
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915456700383232
author Ogbogu, Chukwufumnanya
Narang, Gaurav
Joardar, Biresh Kumar
Doppa, Janardhan Rao
Chakrabarty, Krishnendu
Pande, Partha Pratim
author_facet Ogbogu, Chukwufumnanya
Narang, Gaurav
Joardar, Biresh Kumar
Doppa, Janardhan Rao
Chakrabarty, Krishnendu
Pande, Partha Pratim
contents Processing-In-Memory (PIM) architectures offer a promising approach to accelerate Graph Neural Network (GNN) training and inference. However, various PIM devices such as ReRAM, FeFET, PCM, MRAM, and SRAM exist, with each device offering unique trade-offs in terms of power, latency, area, and non-idealities. A heterogeneous manycore architecture enabled by 3D integration can combine multiple PIM devices on a single platform, to enable energy-efficient and high-performance GNN training. In this work, we propose a 3D heterogeneous PIM-based accelerator for GNN training referred to as HePGA. We leverage the unique characteristics of GNN layers and associated computing kernels to optimize their mapping on to different PIM devices as well as planar tiers. Our experimental analysis shows that HePGA outperforms existing PIM-based architectures by up to 3.8x and 6.8x in energy-efficiency (TOPS/W) and compute efficiency (TOPS/mm2) respectively, without sacrificing the GNN prediction accuracy. Finally, we demonstrate the applicability of HePGA to accelerate inferencing of emerging transformer models.
format Preprint
id arxiv_https___arxiv_org_abs_2508_16011
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle HePGA: A Heterogeneous Processing-in-Memory based GNN Training Accelerator
Ogbogu, Chukwufumnanya
Narang, Gaurav
Joardar, Biresh Kumar
Doppa, Janardhan Rao
Chakrabarty, Krishnendu
Pande, Partha Pratim
Emerging Technologies
Hardware Architecture
Machine Learning
Processing-In-Memory (PIM) architectures offer a promising approach to accelerate Graph Neural Network (GNN) training and inference. However, various PIM devices such as ReRAM, FeFET, PCM, MRAM, and SRAM exist, with each device offering unique trade-offs in terms of power, latency, area, and non-idealities. A heterogeneous manycore architecture enabled by 3D integration can combine multiple PIM devices on a single platform, to enable energy-efficient and high-performance GNN training. In this work, we propose a 3D heterogeneous PIM-based accelerator for GNN training referred to as HePGA. We leverage the unique characteristics of GNN layers and associated computing kernels to optimize their mapping on to different PIM devices as well as planar tiers. Our experimental analysis shows that HePGA outperforms existing PIM-based architectures by up to 3.8x and 6.8x in energy-efficiency (TOPS/W) and compute efficiency (TOPS/mm2) respectively, without sacrificing the GNN prediction accuracy. Finally, we demonstrate the applicability of HePGA to accelerate inferencing of emerging transformer models.
title HePGA: A Heterogeneous Processing-in-Memory based GNN Training Accelerator
topic Emerging Technologies
Hardware Architecture
Machine Learning
url https://arxiv.org/abs/2508.16011