Finding exact hitting set solutions for systems biology applications using heterogeneous GPU clusters
详细信息 查看全文
- 作者：Danilo Carastan-Santos ; danilo.santos@ufabc.edu.br" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae ; Raphael Y. de Camargo raphael.camargo@ufabc.edu.br" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae ; David C. Martins Jr. david.martins@ufabc.edu.br" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae ; Siang W. Song siang.song@ufabc.edu.br" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae ; Luiz C.S. Rozante luiz.rozante@ufabc.edu.br" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae
- 关键词：Hitting set ; GRN inference ; GPU computing ; CUDA ; GPU cluster
- 刊名：Future Generation Computer Systems
- 出版年：2017
- 出版时间：February 2017
- 年：2017
- 卷：67
- 期：Complete
- 页码：418-429
- 全文大小：1462 K
文摘
- class="listitem" id="list_l000005">
- class="label">•
The algorithm works for Hitting Set Problems (HSP) with thousands of variables.
- class="label">•
A preprocessing sorting procedure speedups the HSP algorithm 2.48 times.
- class="label">•
The GPU algorithm is up to 7.8 times faster than a parallel ten-core CPU version.
- class="label">•
The multi-GPU version load-balances the work among devices on heterogeneous GPUs.
- class="label">•
The algorithm on 8 GPUs runs 118.5 times faster than the original in a ten-core CPU.