Parallel Scalability and Efficiency of HEVC Parallelization Approaches

Ching, Chi Bun; Alvarez-Mesa, Mauricio; Juurlink, Ben; Clare, Gordon; Henry, Félix; Pateux, Stéphane; Schierl, Thomas

doi:10.1109/tcsvt.2012.2223056

Cited by 208 publications

(32 citation statements)

References 10 publications

(11 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the CTU level, by initializing the context model for each line, and getting the prediction information from the top-right CTU, each CTU can start encoding immediately after its top-right CTU is encoded, forming a wavefront of CTUs that can be processed in parallel. Compared with slices and tiles, the absence of a hard boundary in WPP leads to significant RD performance improvement while maintaining good parallelism [3].…”

Section: Wavefront Parallel Processing (Wpp)mentioning

confidence: 99%

“…Several parallel processing friendly tools, such as tiles and wavefront parallel processing (WPP) were introduced. Among these, WPP has been reported to achieve a good tradeoff between parallelism and coding performance [3]. Later, several interframe WPP methods were introduced to further improve the parallelism scalability [3] [4][5].…”

Section: Introductionmentioning

confidence: 99%

“…Among these, WPP has been reported to achieve a good tradeoff between parallelism and coding performance [3]. Later, several interframe WPP methods were introduced to further improve the parallelism scalability [3] [4][5]. However, the actual parallelism achieved by the existing methods is still quite limited, e.g.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Novel 3D-WPP algorithms for parallel HEVC encoding

Wen

Quo

Liu

et al. 2016

2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Although wavefront parallel processing (WPP) proposed in the HEVC standard and various inter frame WPP algorithms can achieve comparatively high parallelism, their scalability for its parallelism is still very limited due to various dependencies introduced in spatial and temporal prediction in HEVC. In this paper, we propose three types of 3 Dimensional WPP (3D-WPP) algorithms that can significantly improve the parallelism, while achieving good tradeoffs between implementation complexity, determinism, and rate-distortion (RD) performance. Experimental results show that the proposed algorithms can lead to up to 2.8x speed up compared with existing inter frame WPP methods. While the Simple 3D-WPP and Static 3D-WPP algorithm may introduce an BD rate loss between 0 to 4.9% as compared with existing algorithms, the more complex Dynamic 3D-WPP algorithm achieves better parallelism with virtually no coding performance loss.

show abstract

Section: Wavefront Parallel Processing (Wpp)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Novel 3D-WPP algorithms for parallel HEVC encoding

Wen

Quo

Liu

et al. 2016

2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…However, this improvement in terms of compression efficiency is bound to a significant increase in the computational complexity. Several works about complexity analysis and parallelization strategies for the HEVC standard can be found in the literature [4][5][6]. Many of the parallelization research efforts have been conducted on the HEVC decoding side.…”

Section: Introductionmentioning

confidence: 99%

Frame-Based and Subpicture-Based Parallelization Approaches of the HEVC Video Encoder

et al. 2018

View full text Add to dashboard Cite

Abstract:The most recent video coding standard, High Efficiency Video Coding (HEVC), is able to significantly improve the compression performance at the expense of a huge computational complexity increase with respect to its predecessor, H.264/AVC. Parallel versions of the HEVC encoder may help to reduce the overall encoding time in order to make it more suitable for practical applications. In this work, we study two parallelization strategies. One of them follows a coarse-grain approach, where parallelization is based on frames, and the other one follows a fine-grain approach, where parallelization is performed at subpicture level. Two different frame-based approaches have been developed. The first one only uses MPI and the second one is a hybrid MPI/OpenMP algorithm. An exhaustive experimental test was carried out to study the performance of both approaches in order to find out the best setup in terms of parallel efficiency and coding performance. Both frame-based and subpicture-based approaches are compared under the same hardware platform. Although subpicture-based schemes provide an excellent performance with high-resolution video sequences, scalability is limited by resolution, and the coding performance worsens by increasing the number of processes. Conversely, the proposed frame-based approaches provide the best results with respect to both parallel performance (increasing scalability) and coding performance (not degrading the rate/distortion behavior).

show abstract

“…Yet, its computations still depend on the data and it requires thread synchronizations after each line. Transform based methods do either suffer from lower compression ratios like JPEG 2000, currently widely used in medical applications, or exploit non-local image or inter-image dependencies like in current video coding standards that are thus restricted to moderate parallelization, e. g., using wavefront ordering [10,11].…”

Section: Introductionmentioning

confidence: 99%

Massively parallel lossless compression of medical images using least-squares prediction and arithmetic coding

Weinlich

Rehm

Amon

et al. 2013

2013 IEEE International Conference on Image Processing

View full text Add to dashboard Cite

Medical imaging in hospitals requires fast and efficient image compression to support the clinical work flow and to save costs. Leastsquares autoregressive pixel prediction methods combined with arithmetic coding constitutes the state of the art in lossless image compression. However, a high computational complexity of both prevents the application of respective CPU implementations in practice. We present a massively parallel compression system for medical volume images which runs on graphics cards. Image blocks are processed independently by separate processing threads. After pixel prediction with specialized border treatment, prediction errors are entropy coded with an adaptive binary arithmetic coder. Both steps are designed to match particular demands of the parallel hardware architecture. Comparisons with current image and video coders show efficiency gains of 3.3-13.6% while compression times can be reduced to a few seconds.

show abstract

Parallel Scalability and Efficiency of HEVC Parallelization Approaches

Cited by 208 publications

References 10 publications

Novel 3D-WPP algorithms for parallel HEVC encoding

Novel 3D-WPP algorithms for parallel HEVC encoding

Frame-Based and Subpicture-Based Parallelization Approaches of the HEVC Video Encoder

Massively parallel lossless compression of medical images using least-squares prediction and arithmetic coding

Contact Info

Product

Resources

About