Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FeNOMS: Enhancing Open Modification Spectral Library Search with In-Storage Processing on Ferroelectric NAND (FeNAND) Flash

Sumukh Pinge, Ashkan Moradifirouzabadi, Keming Fan, Prasanna Venkatesan Ravindran, Tanvir H. Pantha, Po-Kai Hsu, Zheyu Li, Weihong Xu, Zihan Xia, Flavio Ponzina, Winston Chern, Taeyoung Song, Priyankka Ravikumar, Mengkun Tian, Lance Fernandes, Huy Tran, Hari Jayasankar, Hang Chen, Chinsung Park, Amrit Garlapati, Kijoon Kim, Jongho Woo, Suhwan Lim, Kwangsoo Kim, Wanki Kim, Daewon Ha, Duygu Kuzum, Shimeng Yu, Sourav Dutta, Asif Khan, Tajana Rosing, Mingu Kang

TL;DR

Mass spectrometry data volumes render open modification spectral library search (OMS) computationally challenging. The authors propose FeNOMS, an in-storage processing framework that leverages 3D Ferroelectric NAND (FeNAND) and hyperdimensional computing to parallelize distance-based OMS, complemented by Dual-Bound Approximate Matching (D-BAM) to tolerate hardware non-idealities. The approach stores high-dimensional vectors in FeNAND, performs in-memory distance checks with multi-row parallelism, and relies on an external accumulator for post-processing and FDR filtering. Results show substantial improvements, including up to 43× speedup and 21× energy efficiency over state-of-the-art 3D NAND methods, while maintaining comparable accuracy, highlighting FeNOMS as a scalable solution for large-scale OMS and potentially other data-search tasks.

Abstract

The rapid expansion of mass spectrometry (MS) data, now exceeding hundreds of terabytes, poses significant challenges for efficient, large-scale library search - a critical component for drug discovery. Traditional processors struggle to handle this data volume efficiently, making in-storage computing (ISP) a promising alternative. This work introduces an ISP architecture leveraging a 3D Ferroelectric NAND (FeNAND) structure, providing significantly higher density, faster speeds, and lower voltage requirements compared to traditional NAND flash. Despite its superior density, the NAND structure has not been widely utilized in ISP applications due to limited throughput associated with row-by-row reads from serially connected cells. To overcome these limitations, we integrate hyperdimensional computing (HDC), a brain-inspired paradigm that enables highly parallel processing with simple operations and strong error tolerance. By combining HDC with the proposed dual-bound approximate matching (D-BAM) distance metric, tailored to the FeNAND structure, we parallelize vector computations to enable efficient MS spectral library search, achieving 43x speedup and 21x higher energy efficiency over state-of-the-art 3D NAND methods, while maintaining comparable accuracy.

FeNOMS: Enhancing Open Modification Spectral Library Search with In-Storage Processing on Ferroelectric NAND (FeNAND) Flash

TL;DR

Mass spectrometry data volumes render open modification spectral library search (OMS) computationally challenging. The authors propose FeNOMS, an in-storage processing framework that leverages 3D Ferroelectric NAND (FeNAND) and hyperdimensional computing to parallelize distance-based OMS, complemented by Dual-Bound Approximate Matching (D-BAM) to tolerate hardware non-idealities. The approach stores high-dimensional vectors in FeNAND, performs in-memory distance checks with multi-row parallelism, and relies on an external accumulator for post-processing and FDR filtering. Results show substantial improvements, including up to 43× speedup and 21× energy efficiency over state-of-the-art 3D NAND methods, while maintaining comparable accuracy, highlighting FeNOMS as a scalable solution for large-scale OMS and potentially other data-search tasks.

Abstract

The rapid expansion of mass spectrometry (MS) data, now exceeding hundreds of terabytes, poses significant challenges for efficient, large-scale library search - a critical component for drug discovery. Traditional processors struggle to handle this data volume efficiently, making in-storage computing (ISP) a promising alternative. This work introduces an ISP architecture leveraging a 3D Ferroelectric NAND (FeNAND) structure, providing significantly higher density, faster speeds, and lower voltage requirements compared to traditional NAND flash. Despite its superior density, the NAND structure has not been widely utilized in ISP applications due to limited throughput associated with row-by-row reads from serially connected cells. To overcome these limitations, we integrate hyperdimensional computing (HDC), a brain-inspired paradigm that enables highly parallel processing with simple operations and strong error tolerance. By combining HDC with the proposed dual-bound approximate matching (D-BAM) distance metric, tailored to the FeNAND structure, we parallelize vector computations to enable efficient MS spectral library search, achieving 43x speedup and 21x higher energy efficiency over state-of-the-art 3D NAND methods, while maintaining comparable accuracy.

Paper Structure

This paper contains 15 sections, 3 equations, 12 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background and Motivation
  3. MS Pipeline & OMS Search
  4. HD Computing for OMS
  5. In-storage processing within 3D NAND architectures
  6. High-Density, Energy-Efficient Ferroelectric NAND
  7. Proposed FeNOMS ISP Acceleration Framework
  8. FeNOMS ISP Overview
  9. Dual-Bound Approximate Matching (D-BAM)
  10. RESULTS
  11. Experimental Setup
  12. Search Quality Analysis
  13. PPA Improvement Analysis
  14. FeNOMS: Design Space Exploration (DSE)
  15. CONCLUSION

Figures (12)

  • Figure 1: Open modification search using HDC.
  • Figure 2: Seven sequential $V_{\text{read}}$ sensing, for precise reading.
  • Figure 3: FeNOMS system overview with data mapping on 3D FeNAND array and dataflow for D-BAM implementation.
  • Figure 4: Sensing with $\text{PF}_3$, optimized for similarity search.
  • Figure 5: D-BAM tolerance-based similarity checks: query $q$ as wordline voltages, reference $V_{\text{TH}}$ in FeNAND string.
  • ...and 7 more figures