Trackable Island-model Genetic Algorithms at Wafer Scale

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Moreno, Matthew Andres, Yang, Connor, Dolson, Emily, Zaman, Luis
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929336491180032
author Moreno, Matthew Andres
Yang, Connor
Dolson, Emily
Zaman, Luis
author_facet Moreno, Matthew Andres
Yang, Connor
Dolson, Emily
Zaman, Luis
contents Emerging ML/AI hardware accelerators, like the 850,000 processor Cerebras Wafer-Scale Engine (WSE), hold great promise to scale up the capabilities of evolutionary computation. However, challenges remain in maintaining visibility into underlying evolutionary processes while efficiently utilizing these platforms' large processor counts. Here, we focus on the problem of extracting phylogenetic information from digital evolution on the WSE platform. We present a tracking-enabled asynchronous island-based genetic algorithm (GA) framework for WSE hardware. Emulated and on-hardware GA benchmarks with a simple tracking-enabled agent model clock upwards of 1 million generations a minute for population sizes reaching 16 million. This pace enables quadrillions of evaluations a day. We validate phylogenetic reconstructions from these trials and demonstrate their suitability for inference of underlying evolutionary conditions. In particular, we demonstrate extraction of clear phylometric signals that differentiate wafer-scale runs with adaptive dynamics enabled versus disabled. Together, these benchmark and validation trials reflect strong potential for highly scalable evolutionary computation that is both efficient and observable. Kernel code implementing the island-model GA supports drop-in customization to support any fixed-length genome content and fitness criteria, allowing it to be leveraged to advance research interests across the community.
format Preprint
id arxiv_https___arxiv_org_abs_2405_03605
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Trackable Island-model Genetic Algorithms at Wafer Scale
Moreno, Matthew Andres
Yang, Connor
Dolson, Emily
Zaman, Luis
Neural and Evolutionary Computing
Distributed, Parallel, and Cluster Computing
Emerging ML/AI hardware accelerators, like the 850,000 processor Cerebras Wafer-Scale Engine (WSE), hold great promise to scale up the capabilities of evolutionary computation. However, challenges remain in maintaining visibility into underlying evolutionary processes while efficiently utilizing these platforms' large processor counts. Here, we focus on the problem of extracting phylogenetic information from digital evolution on the WSE platform. We present a tracking-enabled asynchronous island-based genetic algorithm (GA) framework for WSE hardware. Emulated and on-hardware GA benchmarks with a simple tracking-enabled agent model clock upwards of 1 million generations a minute for population sizes reaching 16 million. This pace enables quadrillions of evaluations a day. We validate phylogenetic reconstructions from these trials and demonstrate their suitability for inference of underlying evolutionary conditions. In particular, we demonstrate extraction of clear phylometric signals that differentiate wafer-scale runs with adaptive dynamics enabled versus disabled. Together, these benchmark and validation trials reflect strong potential for highly scalable evolutionary computation that is both efficient and observable. Kernel code implementing the island-model GA supports drop-in customization to support any fixed-length genome content and fitness criteria, allowing it to be leveraged to advance research interests across the community.
title Trackable Island-model Genetic Algorithms at Wafer Scale
topic Neural and Evolutionary Computing
Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2405.03605