SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry Analysis
Mass spectrometry (MS) is essential for proteomics and metabolomics but faces impending challenges in efficiently processing the vast volumes of data. This article introduces SpecPCM, an in-memory computing (IMC) accelerator designed to achieve substantial improvements in energy and delay efficiency...
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Journal on Exploratory Solid-State Computational Devices and Circuits |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10753646/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832590354312855552 |
|---|---|
| author | Keming Fan Ashkan Moradifirouzabadi Xiangjin Wu Zheyu Li Flavio Ponzina Anton Persson Eric Pop Tajana Rosing Mingu Kang |
| author_facet | Keming Fan Ashkan Moradifirouzabadi Xiangjin Wu Zheyu Li Flavio Ponzina Anton Persson Eric Pop Tajana Rosing Mingu Kang |
| author_sort | Keming Fan |
| collection | DOAJ |
| description | Mass spectrometry (MS) is essential for proteomics and metabolomics but faces impending challenges in efficiently processing the vast volumes of data. This article introduces SpecPCM, an in-memory computing (IMC) accelerator designed to achieve substantial improvements in energy and delay efficiency for both MS spectral clustering and database (DB) search. SpecPCM employs analog processing with low-voltage swing and utilizes recently introduced phase change memory (PCM) devices based on superlattice materials, optimized for low-voltage and low-power programming. Our approach integrates contributions across multiple levels: application, algorithm, circuit, device, and instruction sets. We leverage a robust hyperdimensional computing (HD) algorithm with a novel dimension-packing method and develop specialized hardware for the end-to-end MS pipeline to overcome the nonideal behavior of PCM devices. We further optimize multilevel PCM devices for different tasks by using different materials. We also perform a comprehensive design exploration to improve energy and delay efficiency while maintaining accuracy, exploring various combinations of hardware and software parameters controlled by the instruction set architecture (ISA). SpecPCM, with up to three bits per cell, achieves speedups of up to <inline-formula> <tex-math notation="LaTeX">$82\times $ </tex-math></inline-formula> and <inline-formula> <tex-math notation="LaTeX">$143\times $ </tex-math></inline-formula> for MS clustering and DB search tasks, respectively, along with a four-orders-of-magnitude improvement in energy efficiency compared with state-of-the-art (SoA) CPU/GPU tools. |
| format | Article |
| id | doaj-art-b9983216cb7447a1b7b68b0dc504f965 |
| institution | Kabale University |
| issn | 2329-9231 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal on Exploratory Solid-State Computational Devices and Circuits |
| spelling | doaj-art-b9983216cb7447a1b7b68b0dc504f9652025-01-24T00:02:10ZengIEEEIEEE Journal on Exploratory Solid-State Computational Devices and Circuits2329-92312024-01-011016116910.1109/JXCDC.2024.349883710753646SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry AnalysisKeming Fan0https://orcid.org/0000-0002-6659-9971Ashkan Moradifirouzabadi1https://orcid.org/0009-0007-5112-300XXiangjin Wu2https://orcid.org/0000-0001-8155-1282Zheyu Li3Flavio Ponzina4https://orcid.org/0000-0002-9662-498XAnton Persson5https://orcid.org/0000-0003-0895-096XEric Pop6https://orcid.org/0000-0003-0436-8534Tajana Rosing7https://orcid.org/0000-0002-6954-997XMingu Kang8https://orcid.org/0000-0001-8104-5136Department of Electrical and Computer Engineering, University of California at San Diego, San Diego, CA, USADepartment of Electrical and Computer Engineering, University of California at San Diego, San Diego, CA, USADepartment of Electrical Engineering, Stanford University, Palo Alto, CA, USADepartment of Computer Science and Engineering, University of California at San Diego, San Diego, CA, USADepartment of Computer Science and Engineering, University of California at San Diego, San Diego, CA, USADepartment of Electrical Engineering, Stanford University, Palo Alto, CA, USADepartment of Electrical Engineering, Stanford University, Palo Alto, CA, USADepartment of Computer Science and Engineering, University of California at San Diego, San Diego, CA, USADepartment of Electrical and Computer Engineering, University of California at San Diego, San Diego, CA, USAMass spectrometry (MS) is essential for proteomics and metabolomics but faces impending challenges in efficiently processing the vast volumes of data. This article introduces SpecPCM, an in-memory computing (IMC) accelerator designed to achieve substantial improvements in energy and delay efficiency for both MS spectral clustering and database (DB) search. SpecPCM employs analog processing with low-voltage swing and utilizes recently introduced phase change memory (PCM) devices based on superlattice materials, optimized for low-voltage and low-power programming. Our approach integrates contributions across multiple levels: application, algorithm, circuit, device, and instruction sets. We leverage a robust hyperdimensional computing (HD) algorithm with a novel dimension-packing method and develop specialized hardware for the end-to-end MS pipeline to overcome the nonideal behavior of PCM devices. We further optimize multilevel PCM devices for different tasks by using different materials. We also perform a comprehensive design exploration to improve energy and delay efficiency while maintaining accuracy, exploring various combinations of hardware and software parameters controlled by the instruction set architecture (ISA). SpecPCM, with up to three bits per cell, achieves speedups of up to <inline-formula> <tex-math notation="LaTeX">$82\times $ </tex-math></inline-formula> and <inline-formula> <tex-math notation="LaTeX">$143\times $ </tex-math></inline-formula> for MS clustering and DB search tasks, respectively, along with a four-orders-of-magnitude improvement in energy efficiency compared with state-of-the-art (SoA) CPU/GPU tools.https://ieeexplore.ieee.org/document/10753646/Hardware-software co-designhyperdimensional computing (HD)in-memory computing (IMC)instruction set architecture (ISA)mass spectrometry (MS)phase change memory (PCM) |
| spellingShingle | Keming Fan Ashkan Moradifirouzabadi Xiangjin Wu Zheyu Li Flavio Ponzina Anton Persson Eric Pop Tajana Rosing Mingu Kang SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry Analysis IEEE Journal on Exploratory Solid-State Computational Devices and Circuits Hardware-software co-design hyperdimensional computing (HD) in-memory computing (IMC) instruction set architecture (ISA) mass spectrometry (MS) phase change memory (PCM) |
| title | SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry Analysis |
| title_full | SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry Analysis |
| title_fullStr | SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry Analysis |
| title_full_unstemmed | SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry Analysis |
| title_short | SpecPCM: A Low-Power PCM-Based In-Memory Computing Accelerator for Full-Stack Mass Spectrometry Analysis |
| title_sort | specpcm a low power pcm based in memory computing accelerator for full stack mass spectrometry analysis |
| topic | Hardware-software co-design hyperdimensional computing (HD) in-memory computing (IMC) instruction set architecture (ISA) mass spectrometry (MS) phase change memory (PCM) |
| url | https://ieeexplore.ieee.org/document/10753646/ |
| work_keys_str_mv | AT kemingfan specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT ashkanmoradifirouzabadi specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT xiangjinwu specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT zheyuli specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT flavioponzina specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT antonpersson specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT ericpop specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT tajanarosing specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis AT mingukang specpcmalowpowerpcmbasedinmemorycomputingacceleratorforfullstackmassspectrometryanalysis |