Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI

Jingrong Zhang; Zihao Wang; Yu Chen; Zhiyong Liu; Fa Zhang

doi:10.1007/978-3-319-94968-0_19

Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI

Jingrong Zhang, Zihao Wang, Yu Chen, Zhiyong Liu, Fa Zhang^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Citations (Scopus)

Abstract

In cryo-electron microscopy, RELION has been proven to be a powerful tool for high-resolution reconstruction and has quickly gained its popularity. However, as the data processed in cryoEM is large and the algorithm of RELION is computation-intensive, the refinement procedure of RELION appears quite time-consuming and memory-demanding. These two problems have become major bottlenecks for its usage. Even though there have been efforts on paralleling RELION, the global memory size still may not meet its requirement. Also as by now there is no automatic memory management system on GPU (Graphics Processing Unit), the fragmentation will increase with iteration. Eventually, it would crash the program. In our work, we designed a memory-efficient and stabilizing management system to guarantee the robustness of our program and the efficiency of GPU memory usage. To reduce the memory usage, we developed a novel RELION 2.0 data structure. Also, we proposed a weight calculation parallel algorithm to speedup the calculation. Experiments show that the memory system can avoid memory fragmentation and we can achieve better speedup ratio compared with RELION 2.0.

Original language	English
Title of host publication	Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings
Editors	Fa Zhang, Shihua Zhang, Zhipeng Cai, Pavel Skums
Publisher	Springer Verlag
Pages	205-216
Number of pages	12
ISBN (Print)	9783319949673
DOIs	https://doi.org/10.1007/978-3-319-94968-0_19
Publication status	Published - 2018
Externally published	Yes
Event	14th International Symposium on Bioinformatics Research and Applications, ISBRA 2018 - Beijing, China Duration: 8 Jun 2018 → 11 Jun 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10847 LNBI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th International Symposium on Bioinformatics Research and Applications, ISBRA 2018
Country/Territory	China
City	Beijing
Period	8/06/18 → 11/06/18

Keywords

CUDA
CryoEM
Performance tuning
RELION

Access to Document

10.1007/978-3-319-94968-0_19

Cite this

Zhang, J., Wang, Z., Chen, Y., Liu, Z., & Zhang, F. (2018). Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI. In F. Zhang, S. Zhang, Z. Cai, & P. Skums (Eds.), Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings (pp. 205-216). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10847 LNBI). Springer Verlag. https://doi.org/10.1007/978-3-319-94968-0_19

Zhang, Jingrong ; Wang, Zihao ; Chen, Yu et al. / Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI. Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings. editor / Fa Zhang ; Shihua Zhang ; Zhipeng Cai ; Pavel Skums. Springer Verlag, 2018. pp. 205-216 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{b3aae1e3ac7c4a46baaa245dbeac16ff,

title = "Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI",

abstract = "In cryo-electron microscopy, RELION has been proven to be a powerful tool for high-resolution reconstruction and has quickly gained its popularity. However, as the data processed in cryoEM is large and the algorithm of RELION is computation-intensive, the refinement procedure of RELION appears quite time-consuming and memory-demanding. These two problems have become major bottlenecks for its usage. Even though there have been efforts on paralleling RELION, the global memory size still may not meet its requirement. Also as by now there is no automatic memory management system on GPU (Graphics Processing Unit), the fragmentation will increase with iteration. Eventually, it would crash the program. In our work, we designed a memory-efficient and stabilizing management system to guarantee the robustness of our program and the efficiency of GPU memory usage. To reduce the memory usage, we developed a novel RELION 2.0 data structure. Also, we proposed a weight calculation parallel algorithm to speedup the calculation. Experiments show that the memory system can avoid memory fragmentation and we can achieve better speedup ratio compared with RELION 2.0.",

keywords = "CUDA, CryoEM, Performance tuning, RELION",

author = "Jingrong Zhang and Zihao Wang and Yu Chen and Zhiyong Liu and Fa Zhang",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG, part of Springer Nature 2018.; 14th International Symposium on Bioinformatics Research and Applications, ISBRA 2018 ; Conference date: 08-06-2018 Through 11-06-2018",

year = "2018",

doi = "10.1007/978-3-319-94968-0_19",

language = "English",

isbn = "9783319949673",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "205--216",

editor = "Fa Zhang and Shihua Zhang and Zhipeng Cai and Pavel Skums",

booktitle = "Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings",

address = "Germany",

}

Zhang, J, Wang, Z, Chen, Y, Liu, Z & Zhang, F 2018, Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI. in F Zhang, S Zhang, Z Cai & P Skums (eds), Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10847 LNBI, Springer Verlag, pp. 205-216, 14th International Symposium on Bioinformatics Research and Applications, ISBRA 2018, Beijing, China, 8/06/18. https://doi.org/10.1007/978-3-319-94968-0_19

Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI. / Zhang, Jingrong; Wang, Zihao; Chen, Yu et al.
Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings. ed. / Fa Zhang; Shihua Zhang; Zhipeng Cai; Pavel Skums. Springer Verlag, 2018. p. 205-216 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10847 LNBI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI

AU - Zhang, Jingrong

AU - Wang, Zihao

AU - Chen, Yu

AU - Liu, Zhiyong

AU - Zhang, Fa

N1 - Publisher Copyright: © Springer International Publishing AG, part of Springer Nature 2018.

PY - 2018

Y1 - 2018

N2 - In cryo-electron microscopy, RELION has been proven to be a powerful tool for high-resolution reconstruction and has quickly gained its popularity. However, as the data processed in cryoEM is large and the algorithm of RELION is computation-intensive, the refinement procedure of RELION appears quite time-consuming and memory-demanding. These two problems have become major bottlenecks for its usage. Even though there have been efforts on paralleling RELION, the global memory size still may not meet its requirement. Also as by now there is no automatic memory management system on GPU (Graphics Processing Unit), the fragmentation will increase with iteration. Eventually, it would crash the program. In our work, we designed a memory-efficient and stabilizing management system to guarantee the robustness of our program and the efficiency of GPU memory usage. To reduce the memory usage, we developed a novel RELION 2.0 data structure. Also, we proposed a weight calculation parallel algorithm to speedup the calculation. Experiments show that the memory system can avoid memory fragmentation and we can achieve better speedup ratio compared with RELION 2.0.

AB - In cryo-electron microscopy, RELION has been proven to be a powerful tool for high-resolution reconstruction and has quickly gained its popularity. However, as the data processed in cryoEM is large and the algorithm of RELION is computation-intensive, the refinement procedure of RELION appears quite time-consuming and memory-demanding. These two problems have become major bottlenecks for its usage. Even though there have been efforts on paralleling RELION, the global memory size still may not meet its requirement. Also as by now there is no automatic memory management system on GPU (Graphics Processing Unit), the fragmentation will increase with iteration. Eventually, it would crash the program. In our work, we designed a memory-efficient and stabilizing management system to guarantee the robustness of our program and the efficiency of GPU memory usage. To reduce the memory usage, we developed a novel RELION 2.0 data structure. Also, we proposed a weight calculation parallel algorithm to speedup the calculation. Experiments show that the memory system can avoid memory fragmentation and we can achieve better speedup ratio compared with RELION 2.0.

KW - CUDA

KW - CryoEM

KW - Performance tuning

KW - RELION

UR - http://www.scopus.com/inward/record.url?scp=85050353518&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-94968-0_19

DO - 10.1007/978-3-319-94968-0_19

M3 - Conference contribution

AN - SCOPUS:85050353518

SN - 9783319949673

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 205

EP - 216

BT - Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings

A2 - Zhang, Fa

A2 - Zhang, Shihua

A2 - Cai, Zhipeng

A2 - Skums, Pavel

PB - Springer Verlag

T2 - 14th International Symposium on Bioinformatics Research and Applications, ISBRA 2018

Y2 - 8 June 2018 through 11 June 2018

ER -

Zhang J, Wang Z, Chen Y, Liu Z, Zhang F. Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI. In Zhang F, Zhang S, Cai Z, Skums P, editors, Bioinformatics Research and Applications - 14th International Symposium, ISBRA 2018, Proceedings. Springer Verlag. 2018. p. 205-216. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-94968-0_19

Memory-efficient and stabilizing management system and parallel methods for RELION using CUDA and MPI

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this