Flexible Hierarchical Parallel Processing for AVS3 Video Coding

Hannong Zheng; Yuhuai Zhang; Jian Zhang; Hengyu Man; Xuan Deng; Siwei Ma

doi:10.1007/978-3-031-46311-2_33

Flexible Hierarchical Parallel Processing for AVS3 Video Coding

Hannong Zheng^*
, Yuhuai Zhang
, Jian Zhang
, Hengyu Man
, Xuan Deng
, Siwei Ma

^*Corresponding author for this work

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

Abstract

Video coding standards adopt a series of time-consuming techniques to improve coding performance. Acceleration is vital to the application of real-time video coding. However, conventional frame-level parallelisms usually fail to accelerate effectively, since the coding order is barely adjusted. This motivates more fine-grained parallelisms adopted in industry codecs, while the coding performance loss is inevitably introduced due to the coding restrictions. To benefit parallelism in Random Access (RA) configuration, we propose a novel parallel strategy named Flexible Hierarchical Parallel Processing (FHPP). Based on level order traversal, FHPP achieves cross-GOP and cross-intra-period parallelism in a flexible and hierarchical way. Furthermore, we analyze the upper bound of the FHPP speedup ratio theoretically. Extensive experimental results on the AVS3 standard show that FHPP can achieve an average acceleration of 10 times for 4K sequences under 16 threads, and the maximum acceleration can exceed 12 times, which is significantly better than the traditional frame-level parallelisms.

Original language	English
Title of host publication	Image and Graphics - 12th International Conference, ICIG 2023, Proceedings
Editors	Huchuan Lu, Risheng Liu, Wanli Ouyang, Hui Huang, Jiwen Lu, Jing Dong, Min Xu
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	396-406
Number of pages	11
ISBN (Print)	9783031463105
DOIs	https://doi.org/10.1007/978-3-031-46311-2_33
State	Published - 2023
Externally published	Yes
Event	12th International Conference on Image and Graphics, ICIG 2023 - Nanjing, China Duration: 22 Sep 2023 → 24 Sep 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14357 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	12th International Conference on Image and Graphics, ICIG 2023
Country/Territory	China
City	Nanjing
Period	22/09/23 → 24/09/23

Keywords

Coding order
Frame-level parallelism
Video coding

Access to Document

10.1007/978-3-031-46311-2_33

Cite this

Zheng, H., Zhang, Y., Zhang, J., Man, H., Deng, X., & Ma, S. (2023). Flexible Hierarchical Parallel Processing for AVS3 Video Coding. In H. Lu, R. Liu, W. Ouyang, H. Huang, J. Lu, J. Dong, & M. Xu (Eds.), Image and Graphics - 12th International Conference, ICIG 2023, Proceedings (pp. 396-406). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14357 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-46311-2_33

Zheng, Hannong ; Zhang, Yuhuai ; Zhang, Jian et al. / Flexible Hierarchical Parallel Processing for AVS3 Video Coding. Image and Graphics - 12th International Conference, ICIG 2023, Proceedings. editor / Huchuan Lu ; Risheng Liu ; Wanli Ouyang ; Hui Huang ; Jiwen Lu ; Jing Dong ; Min Xu. Springer Science and Business Media Deutschland GmbH, 2023. pp. 396-406 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{2c6b72e6500b424786c2241e5f7b3d5c,

title = "Flexible Hierarchical Parallel Processing for AVS3 Video Coding",

abstract = "Video coding standards adopt a series of time-consuming techniques to improve coding performance. Acceleration is vital to the application of real-time video coding. However, conventional frame-level parallelisms usually fail to accelerate effectively, since the coding order is barely adjusted. This motivates more fine-grained parallelisms adopted in industry codecs, while the coding performance loss is inevitably introduced due to the coding restrictions. To benefit parallelism in Random Access (RA) configuration, we propose a novel parallel strategy named Flexible Hierarchical Parallel Processing (FHPP). Based on level order traversal, FHPP achieves cross-GOP and cross-intra-period parallelism in a flexible and hierarchical way. Furthermore, we analyze the upper bound of the FHPP speedup ratio theoretically. Extensive experimental results on the AVS3 standard show that FHPP can achieve an average acceleration of 10 times for 4K sequences under 16 threads, and the maximum acceleration can exceed 12 times, which is significantly better than the traditional frame-level parallelisms.",

keywords = "Coding order, Frame-level parallelism, Video coding",

author = "Hannong Zheng and Yuhuai Zhang and Jian Zhang and Hengyu Man and Xuan Deng and Siwei Ma",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.; 12th International Conference on Image and Graphics, ICIG 2023 ; Conference date: 22-09-2023 Through 24-09-2023",

year = "2023",

doi = "10.1007/978-3-031-46311-2\_33",

language = "英语",

isbn = "9783031463105",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "396--406",

editor = "Huchuan Lu and Risheng Liu and Wanli Ouyang and Hui Huang and Jiwen Lu and Jing Dong and Min Xu",

booktitle = "Image and Graphics - 12th International Conference, ICIG 2023, Proceedings",

address = "德国",

}

Zheng, H, Zhang, Y, Zhang, J, Man, H, Deng, X & Ma, S 2023, Flexible Hierarchical Parallel Processing for AVS3 Video Coding. in H Lu, R Liu, W Ouyang, H Huang, J Lu, J Dong & M Xu (eds), Image and Graphics - 12th International Conference, ICIG 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14357 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 396-406, 12th International Conference on Image and Graphics, ICIG 2023, Nanjing, China, 22/09/23. https://doi.org/10.1007/978-3-031-46311-2_33

Flexible Hierarchical Parallel Processing for AVS3 Video Coding. / Zheng, Hannong; Zhang, Yuhuai; Zhang, Jian et al.
Image and Graphics - 12th International Conference, ICIG 2023, Proceedings. ed. / Huchuan Lu; Risheng Liu; Wanli Ouyang; Hui Huang; Jiwen Lu; Jing Dong; Min Xu. Springer Science and Business Media Deutschland GmbH, 2023. p. 396-406 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14357 LNCS).

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

TY - GEN

T1 - Flexible Hierarchical Parallel Processing for AVS3 Video Coding

AU - Zheng, Hannong

AU - Zhang, Yuhuai

AU - Zhang, Jian

AU - Man, Hengyu

AU - Deng, Xuan

AU - Ma, Siwei

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

PY - 2023

Y1 - 2023

N2 - Video coding standards adopt a series of time-consuming techniques to improve coding performance. Acceleration is vital to the application of real-time video coding. However, conventional frame-level parallelisms usually fail to accelerate effectively, since the coding order is barely adjusted. This motivates more fine-grained parallelisms adopted in industry codecs, while the coding performance loss is inevitably introduced due to the coding restrictions. To benefit parallelism in Random Access (RA) configuration, we propose a novel parallel strategy named Flexible Hierarchical Parallel Processing (FHPP). Based on level order traversal, FHPP achieves cross-GOP and cross-intra-period parallelism in a flexible and hierarchical way. Furthermore, we analyze the upper bound of the FHPP speedup ratio theoretically. Extensive experimental results on the AVS3 standard show that FHPP can achieve an average acceleration of 10 times for 4K sequences under 16 threads, and the maximum acceleration can exceed 12 times, which is significantly better than the traditional frame-level parallelisms.

AB - Video coding standards adopt a series of time-consuming techniques to improve coding performance. Acceleration is vital to the application of real-time video coding. However, conventional frame-level parallelisms usually fail to accelerate effectively, since the coding order is barely adjusted. This motivates more fine-grained parallelisms adopted in industry codecs, while the coding performance loss is inevitably introduced due to the coding restrictions. To benefit parallelism in Random Access (RA) configuration, we propose a novel parallel strategy named Flexible Hierarchical Parallel Processing (FHPP). Based on level order traversal, FHPP achieves cross-GOP and cross-intra-period parallelism in a flexible and hierarchical way. Furthermore, we analyze the upper bound of the FHPP speedup ratio theoretically. Extensive experimental results on the AVS3 standard show that FHPP can achieve an average acceleration of 10 times for 4K sequences under 16 threads, and the maximum acceleration can exceed 12 times, which is significantly better than the traditional frame-level parallelisms.

KW - Coding order

KW - Frame-level parallelism

KW - Video coding

UR - https://www.scopus.com/pages/publications/85177475952

U2 - 10.1007/978-3-031-46311-2_33

DO - 10.1007/978-3-031-46311-2_33

M3 - 会议稿件

AN - SCOPUS:85177475952

SN - 9783031463105

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

SP - 396

EP - 406

BT - Image and Graphics - 12th International Conference, ICIG 2023, Proceedings

A2 - Lu, Huchuan

A2 - Liu, Risheng

A2 - Ouyang, Wanli

A2 - Huang, Hui

A2 - Lu, Jiwen

A2 - Dong, Jing

A2 - Xu, Min

PB - Springer Science and Business Media Deutschland GmbH

T2 - 12th International Conference on Image and Graphics, ICIG 2023

Y2 - 22 September 2023 through 24 September 2023

ER -

Zheng H, Zhang Y, Zhang J, Man H, Deng X, Ma S. Flexible Hierarchical Parallel Processing for AVS3 Video Coding. In Lu H, Liu R, Ouyang W, Huang H, Lu J, Dong J, Xu M, editors, Image and Graphics - 12th International Conference, ICIG 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 396-406. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-46311-2_33