GPU-acceleration of the distributed-memory database peptide search of mass spectrometry data.

GPU-acceleration of the distributed-memory database peptide search of mass spectrometry data.
复制标题

DOI:
10.1038/s41598-023-43033-w
复制
发表时间:
2023-10-31
期刊:
影响因子:
4.6
通讯作者:
Saeed, Fahad
Saeed, Fahad
中科院分区:
综合性期刊3区
文献类型:
--
作者:
Haseeb, Muhammad;Saeed, Fahad

文献摘要

参考文献

相似文献

数据库肽搜索是从质谱 (MS) 数据中识别肽的主要计算技术。图形处理单元(GPU)计算现在在当前一代高性能计算(HPC)系统中无处不在,但其在数据库肽搜索领域的应用仍然有限。部分原因是现有 GPU 加速方法中使用了次优算法,导致硬件利用率明显低下。在本文中,我们设计并实现了一个名为 GiCOPS 的新型 CPU-GPU HPC 框架,用于在超级计算机上对现代数据库肽搜索算法进行高效且完整的 GPU 加速。我们的实验表明,在足够大的实验规模下,GiCOPS 比其仅使用 CPU 的前身 HiCOPS 的速度提高了 1.2 到 5 倍,比几种现有的基于 GPU 的数据库搜索算法提高了 10 倍以上。我们使用 Roofline 模型进一步评估和优化框架的性能,并报告多个指标的近乎最佳结果,包括每秒计算量、占用率、内存工作负载、分支效率和共享内存性能。最后,我们针对复杂的整数和内存限制的算法管道提出的 CPU-GPU 方法和优化也可以扩展,以加速现有和未来的肽识别算法。 GiCOPS 现已与我们的 HPC 框架 HiCOPS 集成,可从以下网址获取:https://github.com/pcdslab/gicops。
Database peptide search is the primary computational technique for identifying peptides from the mass spectrometry (MS) data. Graphical Processing Units (GPU) computing is now ubiquitous in the current-generation of high-performance computing (HPC) systems, yet its application in the database peptide search domain remains limited. Part of the reason is the use of sub-optimal algorithms in the existing GPU-accelerated methods resulting in significantly inefficient hardware utilization. In this paper, we design and implement a new-age CPU-GPU HPC framework, called GiCOPS, for efficient and complete GPU-acceleration of the modern database peptide search algorithms on supercomputers. Our experimentation shows that the GiCOPS exhibits between 1.2 to 5 speed improvement over its CPU-only predecessor, HiCOPS, and over 10 improvement over several existing GPU-based database search algorithms for sufficiently large experiment sizes. We further assess and optimize the performance of our framework using the Roofline Model and report near-optimal results for several metrics including computations per second, occupancy rate, memory workload, branch efficiency and shared memory performance. Finally, the CPU-GPU methods and optimizations proposed in our work for complex integer- and memory-bounded algorithmic pipelines can also be extended to accelerate the existing and future peptide identification algorithms. GiCOPS is now integrated with our umbrella HPC framework HiCOPS and is available at: https://github.com/pcdslab/gicops.
DOI: 10.1016/j.cels.2018.08.009
发表时间: 2018-10-24
期刊: Cell systems
影响因子: 9.3
作者:
Beyter D;Lin MS;Yu Y;Pieper R;Bafna V
通讯作者: Bafna V
DOI: 10.1038/s43588-021-00113-z
发表时间: 2021-08
期刊: Nature computational science
影响因子: --
作者:
Haseeb M;Saeed F
通讯作者: Saeed F
DOI: 10.1038/s41587-019-0067-5
发表时间: 2019-04-01
影响因子: 46.9
作者:
Devabhaktuni, Arun;Lin, Sarah;Elias, Joshua E.
通讯作者: Elias, Joshua E.
DOI: 10.1093/bioinformatics/bth092
发表时间: 2004-06-12
期刊: BIOINFORMATICS
影响因子: 5.8
作者:
Craig, R;Beavis, RC
通讯作者: Beavis, RC
DOI: 10.1021/ac0258709
发表时间: 2003-02-15
影响因子: 7.4
作者:
Fenyö, D;Beavis, RC
通讯作者: Beavis, RC