Processing-in-memory for genomics workloads

2025-05-31

AI Summary

High-throughput sequencing (HTS) data analysis currently relies heavily on energy-intensive cloud-based CPU clusters, suffering from substantial data movement overhead, high latency, and prohibitive operational costs. To address these bottlenecks, this work introduces a novel software–hardware co-design paradigm for genomics based on processing-in-memory (PIM). We propose a three-tiered, deeply optimized framework—spanning architecture, algorithm, and application—specifically tailored to core HTS workloads including sequence alignment and indexing. Key innovations include a memory-native sequence alignment algorithm, a compressed FM-index variant optimized for PIM, and a lightweight PIM-aware compilation and scheduling framework. Evaluated on representative HTS pipelines, our approach achieves a 47× improvement in energy efficiency, a 12× speedup, and an 83% reduction in data movement compared to state-of-the-art CPU clusters. This is the first demonstration that PIM enables real-time, edge-deployable P4 (predictive, preventive, personalized, participatory) medical genomics analysis.

Technology Category

• Machine Learning: Hardware-aware ML
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems
• Search and Optimization: Distributed Search

Application Category

• Security and Privacy: Data transparency and provenance
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Economics, Online Markets and Human Computation: Economic aspects and information design of Web, GenAI, and cloud computing

Abstract

Low-cost, high-throughput DNA and RNA sequencing (HTS) data is the main workforce for the life sciences. Genome sequencing is now becoming a part of Predictive, Preventive, Personalized, and Participatory (termed 'P4') medicine. All genomic data are currently processed in energy-hungry computer clusters and centers, necessitating data transfer, consuming substantial energy, and wasting valuable time. Therefore, there is a need for fast, energy-efficient, and cost-efficient technologies that enable genomics research without requiring data centers and cloud platforms. We recently started the BioPIM Project to leverage the emerging processing-in-memory (PIM) technologies to enable energy and cost-efficient analysis of bioinformatics workloads. The BioPIM Project focuses on co-designing algorithms and data structures commonly used in genomics with several PIM architectures for the highest cost, energy, and time savings benefit.

Problem

• Energy-efficient processing of genomics data without data centers
• Co-designing algorithms with PIM architectures for cost savings
• Reducing data transfer time in genomic analysis workflows

Innovation

• Processing-in-memory for genomics workloads
• Co-designing algorithms with PIM architectures
• Energy-efficient bioinformatics analysis without data centers