FloatSOM: GPU-Accelerated, Distributed, Topology-Flexible Self-Organizing Maps

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Tony, Klamt, Sarah, Turner, Katherine, Brustle, Anne, Marsh-Wakefield, Felix, Putri, Givanna
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910177413824512
author Xu, Tony
Klamt, Sarah
Turner, Katherine
Brustle, Anne
Marsh-Wakefield, Felix
Putri, Givanna
author_facet Xu, Tony
Klamt, Sarah
Turner, Katherine
Brustle, Anne
Marsh-Wakefield, Felix
Putri, Givanna
contents GPU-accelerated Self-Organizing Map (SOM) implementations are among the most competitive options for large-scale SOM analysis, but growing dataset sizes increasingly challenge their practical use because workloads no longer fit cleanly within device-memory limits. We introduce FloatSOM, a SOM framework for scalable training and deployment that supports multi-GPU execution, out-of-memory disk-backed streaming, and novel topologies beyond regular lattices. We evaluate FloatSOM on 14 synthetic and real benchmark datasets together with controlled speed scaling benchmarks, and show that these improved topologies, combined with topology-aware hyperparameter fine-tuning, yield lower quantization error than current state-of-the-art SOM baselines. FloatSOM also sustains this performance at large scale with high-throughput distributed execution; in the largest benchmark, it trains a 1024-node SOM network on 1,000,000,000 samples with 50 features in 6.16 minutes on 8 GPUs across two separate high-performance-computing nodes.
format Preprint
id arxiv_https___arxiv_org_abs_2604_26555
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle FloatSOM: GPU-Accelerated, Distributed, Topology-Flexible Self-Organizing Maps
Xu, Tony
Klamt, Sarah
Turner, Katherine
Brustle, Anne
Marsh-Wakefield, Felix
Putri, Givanna
Distributed, Parallel, and Cluster Computing
Machine Learning
GPU-accelerated Self-Organizing Map (SOM) implementations are among the most competitive options for large-scale SOM analysis, but growing dataset sizes increasingly challenge their practical use because workloads no longer fit cleanly within device-memory limits. We introduce FloatSOM, a SOM framework for scalable training and deployment that supports multi-GPU execution, out-of-memory disk-backed streaming, and novel topologies beyond regular lattices. We evaluate FloatSOM on 14 synthetic and real benchmark datasets together with controlled speed scaling benchmarks, and show that these improved topologies, combined with topology-aware hyperparameter fine-tuning, yield lower quantization error than current state-of-the-art SOM baselines. FloatSOM also sustains this performance at large scale with high-throughput distributed execution; in the largest benchmark, it trains a 1024-node SOM network on 1,000,000,000 samples with 50 features in 6.16 minutes on 8 GPUs across two separate high-performance-computing nodes.
title FloatSOM: GPU-Accelerated, Distributed, Topology-Flexible Self-Organizing Maps
topic Distributed, Parallel, and Cluster Computing
Machine Learning
url https://arxiv.org/abs/2604.26555