Optimizing Long-Read Sequence Alignment on a CPU-DSPs Heterogeneous Processor
IEEE Transactions on Computers · 2026-09 · 75 (9) · 3326-3339
DOI: 10.1109/tc.2026.3709833 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Computers, 2026-09. Verfasst von Xinjie An und 6 weiteren Autor:innen. Die Arbeit nennt 35 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Science and Technology Innovation Program of Hunan Province.
Abstract (Originalsprache)
Sequence alignment constitutes the fundamental process of mapping sequencing reads to reference genomes, forming the foundation for numerous genomic analyses. However, long-read sequences produced by third-generation sequencing technologies impose significant computational burdens on alignment. Heterogeneous CPU–DSPs processors offer a promising platform for accelerating this task. However, aligning their architectural features with the computational patterns of sequence alignment remains a significant challenge. In this paper, we present DSPaligner, a novel long-read sequence alignment tool tailored for CPU–DSPs heterogeneous processors. DSPaligner features a collaborative CPU–DSPs execution model and a three-tier parallelism scheme through vectorization, multi-threading and multi-processing. Besides, we employ architecture-aware optimizations to address the four critical challenges: (1) a hierarchical memory management scheme tailored to the DSP memory hierarchy; (2) a read/write window mechanism that reduces data transfer overhead; (3) a dependency elimination strategy based on coordinate transformation; and (4) a double-buffering pipeline for computation-memory overlap. Experiments on the FT-M7032 CPU-DSPs processor show that DSPaligner obtained 63-73 $\boldsymbol{\times}$ × speedup over the baseline when utilizing eight DSP cores within a single cluster, establishing new performance benchmarks for biological sequence alignment on heterogeneous architectures.
Förderung
- National Natural Science Foundation of China
- Science and Technology Innovation Program of Hunan Province
Zitieren
APA
An, X., Guo, Y., Guo, Y., Tang, T., Yang, C., Liao, X., & Cui, Y. (2026). Optimizing Long-Read Sequence Alignment on a CPU-DSPs Heterogeneous Processor. IEEE Transactions on Computers, 75(9), 3326-3339. https://doi.org/10.1109/tc.2026.3709833
Vancouver
An X, Guo Y, Guo Y, Tang T, Yang C, Liao X, et al. Optimizing Long-Read Sequence Alignment on a CPU-DSPs Heterogeneous Processor. IEEE Transactions on Computers. 2026;75(9):3326-3339. doi:10.1109/tc.2026.3709833
BibTeX
@article{An2026,
author = {An, Xinjie and Guo, Yifei and Guo, Yufei and Tang, Tao and Yang, Canqun and Liao, Xiangke and Cui, Yingbo},
title = {Optimizing Long-Read Sequence Alignment on a CPU-DSPs Heterogeneous Processor},
journal = {IEEE Transactions on Computers},
year = {2026},
volume = {75},
number = {9},
pages = {3326--3339},
doi = {10.1109/tc.2026.3709833}
}
EUVIMED