European MPI Users Group Meeting最新文献

筛选
英文 中文
Energy estimation for MPI broadcasting algorithms in large scale HPC systems 大规模高性能计算系统中MPI广播算法的能量估计
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488565
M. Diouri, Olivier Glück, J. Mignot, L. Lefèvre
{"title":"Energy estimation for MPI broadcasting algorithms in large scale HPC systems","authors":"M. Diouri, Olivier Glück, J. Mignot, L. Lefèvre","doi":"10.1145/2488551.2488565","DOIUrl":"https://doi.org/10.1145/2488551.2488565","url":null,"abstract":"Future supercomputers will gather hundreds of millions of communicating cores. The movement of data in such systems will be very energy consuming. We address in this paper the issue of energy consumption of data broadcasting in such large scale systems. To this end, we propose a framework to estimate the energy consumed by different MPI broadcasting algorithms for various execution settings. Validation results show that our estimations are highly accurate and allow to select the least consuming broadcasting algorithm.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"135 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127596684","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 10
Parallel strategies for an inverse docking method 一种逆对接方法的并行策略
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488584
R. Vasseur, S. Baud, L. Steffenel, X. Vigouroux, L. Martiny, M. Krajecki, M. Dauchez
{"title":"Parallel strategies for an inverse docking method","authors":"R. Vasseur, S. Baud, L. Steffenel, X. Vigouroux, L. Martiny, M. Krajecki, M. Dauchez","doi":"10.1145/2488551.2488584","DOIUrl":"https://doi.org/10.1145/2488551.2488584","url":null,"abstract":"Molecular docking is a widely used computational technique that allows studying structure-based interactions complexes between biological objects at the molecular scale. The purpose of the current work is to develop a set of tools that allows performing inverse docking, i.e., to test at a large scale a chemical ligand on a large dataset of proteins, which has several applications on the field of drug research. We developed different strategies to parallelize/distribute the docking procedure, as a way to efficiently exploit the computational performance of multi-core and multi-machine (cluster) environments. The experiments conducted to compare these different strategies encourage the search for decomposing strategies as a way to improve the execution of inverse docking.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"113 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132717785","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 4
A hierarchical parallel storage system based on distributed memory for large scale systems 面向大规模系统的基于分布式存储器的分层并行存储系统
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488598
Francisco Rodrigo Duro, Francisco Javier García Blas, J. Carretero
{"title":"A hierarchical parallel storage system based on distributed memory for large scale systems","authors":"Francisco Rodrigo Duro, Francisco Javier García Blas, J. Carretero","doi":"10.1145/2488551.2488598","DOIUrl":"https://doi.org/10.1145/2488551.2488598","url":null,"abstract":"This paper presents the design and implementation of a storage system for high performance systems based on a multiple level I/O caching architecture. The solution relies on Memcached as a parallel storage system, preserving its powerful capacities such as transparency, quick deployment, and scalability. The designed parallel storage system targets to reduce the I/O latency in data-intensive high performance applications. The proposed solution consists of a user-level library and extended Memcached servers. The solution aims to be hierarchical by deploying Memcached-based I/O servers across all the infrastructure data path. Our experiments demonstrate that our solution is up to 40% faster than PVFS2.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134598539","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 21
McMPI: a managed-code MPI library in pure C# McMPI:一个纯c#的托管代码MPI库
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488572
Daniel J. Holmes, S. Booth
{"title":"McMPI: a managed-code MPI library in pure C#","authors":"Daniel J. Holmes, S. Booth","doi":"10.1145/2488551.2488572","DOIUrl":"https://doi.org/10.1145/2488551.2488572","url":null,"abstract":"This paper presents McMPI, an entirely new MPI library written in C# using only safe managed-code, and performance results from low-level benchmarks demonstrating ping-pong latency and bandwidth comparable with MS-MPI and MPICH2.\u0000 McMPI enables all .Net languages to use MPI messaging without introducing a dependency on unsafe non-managed code, e.g. an existing MPI library. It also takes advantage of .Net thread support to improve intra-node latency.\u0000 This paper also discusses support for multiple threads in McMPI and proposes an extension to the MPI Standard that resolves current ambiguities relating to hosting multiple MPI processes in a single operating system process.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131723800","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 4
Improving MPI applications with a new MPI_Info and the use of the memoization 使用新的MPI_Info和记忆的使用改进MPI应用程序
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488554
A. Calderón, J. Carretero, Félix García Carballeira, Javier Fernández, Daniel Higuero, Borja Bergua
{"title":"Improving MPI applications with a new MPI_Info and the use of the memoization","authors":"A. Calderón, J. Carretero, Félix García Carballeira, Javier Fernández, Daniel Higuero, Borja Bergua","doi":"10.1145/2488551.2488554","DOIUrl":"https://doi.org/10.1145/2488551.2488554","url":null,"abstract":"The MPI forum is actively working for a better MPI standard. The results are the new version 3 of the MPI standard, and the efforts for the incoming MPI 3.1/4.0. The technological changes provide many opportunities for improvements and new ideas. This paper introduces two main contributions in this direction: (1) how to improve the MPI_Info object implementation, and (2) a new way of using the former improved MPI_Info object as a storage solution.\u0000 The MPI_Info object [1] is described as an unordered set of key-value pairs (both key and value are strings, and keys are unique). And it is implemented as a linked list in the two major MPI implementations available.\u0000 We propose a new MPI_Info object implementation that (1) is based on the use of hash tables (what improves the overall performance), and (2) abstracts the underlying hash table infrastructure (what facilitates the usage of the most appropriated solution). Our proposal opens the possibility of extending the use of the MPI_Info object as a shared storage solution among MPI processes. To demonstrate the capabilities of our proposal, we explore the utilization of the memoization technique on MPI applications in order to improve the execution performance.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129141391","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 1
On the performance of concurrent transfers in collective algorithms 集体算法中并发传输性能的研究
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488601
Juan A. Rico-Gallego, Juan Carlos Díaz Martín
{"title":"On the performance of concurrent transfers in collective algorithms","authors":"Juan A. Rico-Gallego, Juan Carlos Díaz Martín","doi":"10.1145/2488551.2488601","DOIUrl":"https://doi.org/10.1145/2488551.2488601","url":null,"abstract":"Inter- and intra-machine MPI collective operations in current multicore clusters are essentially different, and therefore their performance modelling ask for different approaches. Inside a multicore each individual message transmission in a collective operation flow in parallel with others, but sharing the bandwidth of the main memory channel. Current models ignore this issue, making errors like giving the same cost estimation to quite different collective algorithms. We outline a new cost model focused on shared channels.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121775214","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 2
Efficient and truly passive MPI-3 RMA using InfiniBand atomics 高效和真正无源MPI-3 RMA使用InfiniBand原子
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488573
Mingzhe Li, S. Potluri, Khaled Hamidouche, Jithin Jose, D. Panda
{"title":"Efficient and truly passive MPI-3 RMA using InfiniBand atomics","authors":"Mingzhe Li, S. Potluri, Khaled Hamidouche, Jithin Jose, D. Panda","doi":"10.1145/2488551.2488573","DOIUrl":"https://doi.org/10.1145/2488551.2488573","url":null,"abstract":"Multi/many-core architectures offer high compute density on modern supercomputing clusters. It is critical for applications to minimize communication and synchronization overheads to achieve peak performance. MPI offers one-sided communication semantics that are aimed at enabling this. In this paper, we propose a novel design for implementing truly passive shared and exclusive MPI_Win_lock/unlock using InfiniBand atomics. We address limitations in designs published earlier. We also present the impact of our design on MPI_Win_lock all introduced in MPI-3. We demonstrate superior overlap compared to existing two-sided implementations. Using Splash LU kernel, proposed design delivers upto 49% performance improvement compared to existing designs.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123239569","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 9
Triangulating molecular surfaces on multiple GPUs 在多个gpu上对分子表面进行三角测量
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488582
Sérgio Dias, A. Gomes
{"title":"Triangulating molecular surfaces on multiple GPUs","authors":"Sérgio Dias, A. Gomes","doi":"10.1145/2488551.2488582","DOIUrl":"https://doi.org/10.1145/2488551.2488582","url":null,"abstract":"Current GPU-based workstations are inadequate to triangulate and rendering large molecular datasets with thousands and hundreds of thousands, not to say millions, of atoms. The problem is not so the lack of processing power, but the memory limitations of current GPU graphics cards. For example, the NVidia GeForce GTX 590 graphics card comes with two 1.5GB GPUs. We tackle here this problem using a OpenMP-CUDA solution that runs on a loosely-coupled GPU cluster. Basically, we propose a fast, scalable, parallel triangulation algorithm for molecular surfaces that takes advantage of multicore processors of CPUs and GPUs of modern hardware architectures, where each CPU core works as the master of a single GPU, being the processing burden distributed over the CPU cores available in a single computer or a cluster. As much as we know, this is the first marching cubes algorithm that triangulates molecular surfaces on multiple GPUs using CUDA and OpenMP.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123379253","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 7
Run-time creation of the turbulent channel flow database by an HPC simulation using MPI-DB 利用MPI-DB进行HPC仿真,在运行时创建湍流通道流数据库
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488574
Jason Graham, E. Givelberg, Kalin Kanov
{"title":"Run-time creation of the turbulent channel flow database by an HPC simulation using MPI-DB","authors":"Jason Graham, E. Givelberg, Kalin Kanov","doi":"10.1145/2488551.2488574","DOIUrl":"https://doi.org/10.1145/2488551.2488574","url":null,"abstract":"We demonstrate a method based on MPI client-server implementation for storing the output of computations directly into the database. Our method automates the previously used inefficient ingestion process which required development of special tools for each simulation. In large-scale channel flow simulation experiments we were able to ingest the output data sets in real time, without delaying the simulation process. This was accomplished by building a Fortran interface to the MPI-DB software library and using it within the simulation code. The channel flow simulation data set will be exposed for analysis to researchers using the JHU Public Turbulence Database [7].","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114619877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 3
A dynamic pipeline for RNA sequencing on multicore processors 在多核处理器上进行RNA测序的动态流水线
European MPI Users Group Meeting Pub Date : 2013-09-15 DOI: 10.1145/2488551.2488581
Héctor Martínez, Joaquín Tárraga, Ignacio Medina, S. Barrachina, M. Castillo, J. Dopazo, E. S. Quintana‐Ortí
{"title":"A dynamic pipeline for RNA sequencing on multicore processors","authors":"Héctor Martínez, Joaquín Tárraga, Ignacio Medina, S. Barrachina, M. Castillo, J. Dopazo, E. S. Quintana‐Ortí","doi":"10.1145/2488551.2488581","DOIUrl":"https://doi.org/10.1145/2488551.2488581","url":null,"abstract":"We present a concurrent algorithm for mapping short and long RNA sequences on multicore processors. Our solution processes the data, initially stored on disk, in batches of reads which are passed between the consecutive stages of a pipeline. A major operational reorganization of the original static pipeline, combined with a complete reimplementation based on POSIX threads, renders a dissociated execution between threads and stages/task types, so that threads can compute any type of pending task resulting in a dynamic pipeline. The experiments on a multicore platform reveal that this reorganization yields significantly higher performance, specially for architectures equipped with a small to moderate number of cores.\u0000 As an additional contribution, our experiments also reveal that the use of 16-nucleotide (nt) seeds during the one of the stages of the pipeline, instead of the 15-nt length that was proposed originally, yields a remarkable reduction in the execution time of the global alignment process while maintaining the sensitivity of the algorithm.","PeriodicalId":243229,"journal":{"name":"European MPI Users Group Meeting","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125283825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 8
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
相关产品
×
本文献相关产品
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信
小红书