Research Article
Enhancing Sequence Alignment Efficiency Through Concurrent Utilization of Multiple Arm Processors in a Sequential Processing Framework
@INPROCEEDINGS{10.1007/978-3-031-80713-8_16, author={Yunzi Dai and Liwei Liu and Zhuochen Yang and Shaolong Chen}, title={Enhancing Sequence Alignment Efficiency Through Concurrent Utilization of Multiple Arm Processors in a Sequential Processing Framework}, proceedings={Data Information in Online Environments. 4th EAI International Conference, DIONE 2023, Nanchang, China, November 25--27, 2023, Proceedings}, proceedings_a={DIONE}, year={2025}, month={2}, keywords={HPC Embedded system Arm processor Sequence alignment Variant analysis}, doi={10.1007/978-3-031-80713-8_16} }- Yunzi Dai
Liwei Liu
Zhuochen Yang
Shaolong Chen
Year: 2025
Enhancing Sequence Alignment Efficiency Through Concurrent Utilization of Multiple Arm Processors in a Sequential Processing Framework
DIONE
Springer
DOI: 10.1007/978-3-031-80713-8_16
Abstract
High performance computing (HPC) solutions traditionally meet the intense computational demands inherent in genome processing. Components of genome processing have been implemented on GPUs, FPGAs, and ASICs. However, they are primarily used as coprocessors for processor servers rather than as independent running systems. The embedded systems with Arm processors have garnered increasing attention over the years. This study introduces an intelligent technique for short read genome alignment, leveraging advanced Arm-based processors. Our novel system integrates a sequential workflow of Arm-based processors, an intelligent mechanism for partitioning the reference genome, a timer-based system for detecting alignment, and a technique for optimizing memory. This results in accelerated alignment with efficient resource utilization and minimization of unproductive searches. The main focus of this study is the implementation of a workflow that reduces memory access and per-processor footprint, providing a revolutionary approach to aligning short read genomes. Through testing millions of simulated sequences, our system significantly improved both alignment speed and result accuracy.
