Exploration of Alternative GPU Implementations of the Pair-HMMs Forward Algorithm

Shanshan Ren; Koen Bertels; Zaid Al-Ars

doi:10.1109/BIBM.2016.7822645

Exploration of Alternative GPU Implementations of the Pair-HMMs Forward Algorithm

Shanshan Ren, Koen Bertels, Zaid Al-Ars

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific

9 Citations (Scopus)

62 Downloads (Pure)

Abstract

In order to handle the massive raw data generated by next generation sequencing (NGS) platforms, GPUs are widely used by many genetic analysis tools to speed up the used algorithms. In this paper, we use GPUs to accelerate the pair-HMMs forward algorithm, which is used to calculate the overall alignment probability in many genomics analysis tools. We firstly evaluate two different implementation methods to accelerate the pair-HMMs forward algorithm according to their effectiveness on GPU platforms. Based on these two methods, we present several implementations of the pair-HMMs forward algorithm.
We execute these implementations on the NVIDIA Tesla K40 card using different datasets to compare the performance. Experimental results show that the intra-task implementation has the highest throughput in most cases, achieving pure computational throughput as high as 23.56 GCUPS for synthetic datasets.
On a real dataset, the inter-task implementation achieves 4.82x speedup compared with a vectorized implementation executed on a 20-core POWER8 system.

Original language	English
Title of host publication	Proceedings 3rd International Workshop on High Performance Computing on Bioinformatics
Pages	1-8
Number of pages	8
DOIs	https://doi.org/10.1109/BIBM.2016.7822645
Publication status	Published - 2016
Event	The 3rd International Workshop on High Performance Computing on Bioinformatics (HPCB 2016) - Shenzhen, China Duration: 15 Dec 2016 → 18 Dec 2016

Workshop

Workshop	The 3rd International Workshop on High Performance Computing on Bioinformatics (HPCB 2016)
Abbreviated title	HPCB 2016
Country/Territory	China
City	Shenzhen
Period	15/12/16 → 18/12/16

Access to Document

10.1109/BIBM.2016.7822645

10238082Accepted author manuscript, 381 KBLicence: Unspecified

Cite this

@inproceedings{e17032897f504cef9e0d5621a66918ed,

title = "Exploration of Alternative GPU Implementations of the Pair-HMMs Forward Algorithm",

abstract = "In order to handle the massive raw data generated by next generation sequencing (NGS) platforms, GPUs are widely used by many genetic analysis tools to speed up the used algorithms. In this paper, we use GPUs to accelerate the pair-HMMs forward algorithm, which is used to calculate the overall alignment probability in many genomics analysis tools. We firstly evaluate two different implementation methods to accelerate the pair-HMMs forward algorithm according to their effectiveness on GPU platforms. Based on these two methods, we present several implementations of the pair-HMMs forward algorithm.We execute these implementations on the NVIDIA Tesla K40 card using different datasets to compare the performance. Experimental results show that the intra-task implementation has the highest throughput in most cases, achieving pure computational throughput as high as 23.56 GCUPS for synthetic datasets.On a real dataset, the inter-task implementation achieves 4.82x speedup compared with a vectorized implementation executed on a 20-core POWER8 system.",

author = "Shanshan Ren and Koen Bertels and Zaid Al-Ars",

year = "2016",

doi = "10.1109/BIBM.2016.7822645",

language = "English",

pages = "1--8",

booktitle = "Proceedings 3rd International Workshop on High Performance Computing on Bioinformatics",

note = "The 3rd International Workshop on High Performance Computing on Bioinformatics (HPCB 2016), HPCB 2016 ; Conference date: 15-12-2016 Through 18-12-2016",

}

Ren, S, Bertels, K & Al-Ars, Z 2016, Exploration of Alternative GPU Implementations of the Pair-HMMs Forward Algorithm. in Proceedings 3rd International Workshop on High Performance Computing on Bioinformatics. pp. 1-8, The 3rd International Workshop on High Performance Computing on Bioinformatics (HPCB 2016), Shenzhen, China, 15/12/16. https://doi.org/10.1109/BIBM.2016.7822645

TY - GEN

T1 - Exploration of Alternative GPU Implementations of the Pair-HMMs Forward Algorithm

AU - Ren, Shanshan

AU - Bertels, Koen

AU - Al-Ars, Zaid

PY - 2016

Y1 - 2016

N2 - In order to handle the massive raw data generated by next generation sequencing (NGS) platforms, GPUs are widely used by many genetic analysis tools to speed up the used algorithms. In this paper, we use GPUs to accelerate the pair-HMMs forward algorithm, which is used to calculate the overall alignment probability in many genomics analysis tools. We firstly evaluate two different implementation methods to accelerate the pair-HMMs forward algorithm according to their effectiveness on GPU platforms. Based on these two methods, we present several implementations of the pair-HMMs forward algorithm.We execute these implementations on the NVIDIA Tesla K40 card using different datasets to compare the performance. Experimental results show that the intra-task implementation has the highest throughput in most cases, achieving pure computational throughput as high as 23.56 GCUPS for synthetic datasets.On a real dataset, the inter-task implementation achieves 4.82x speedup compared with a vectorized implementation executed on a 20-core POWER8 system.

AB - In order to handle the massive raw data generated by next generation sequencing (NGS) platforms, GPUs are widely used by many genetic analysis tools to speed up the used algorithms. In this paper, we use GPUs to accelerate the pair-HMMs forward algorithm, which is used to calculate the overall alignment probability in many genomics analysis tools. We firstly evaluate two different implementation methods to accelerate the pair-HMMs forward algorithm according to their effectiveness on GPU platforms. Based on these two methods, we present several implementations of the pair-HMMs forward algorithm.We execute these implementations on the NVIDIA Tesla K40 card using different datasets to compare the performance. Experimental results show that the intra-task implementation has the highest throughput in most cases, achieving pure computational throughput as high as 23.56 GCUPS for synthetic datasets.On a real dataset, the inter-task implementation achieves 4.82x speedup compared with a vectorized implementation executed on a 20-core POWER8 system.

U2 - 10.1109/BIBM.2016.7822645

DO - 10.1109/BIBM.2016.7822645

M3 - Conference contribution

SP - 1

EP - 8

BT - Proceedings 3rd International Workshop on High Performance Computing on Bioinformatics

T2 - The 3rd International Workshop on High Performance Computing on Bioinformatics (HPCB 2016)

Y2 - 15 December 2016 through 18 December 2016

ER -

Exploration of Alternative GPU Implementations of the Pair-HMMs Forward Algorithm

Abstract

Workshop

Access to Document

Fingerprint

Cite this