Cargando…

Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning

<!--HTML--><p align="justify"> Physics experiments nowadays produce tremendous amounts of data that require sophisticated analyses in order to gain new insights. At such large scale, scientists are facing non-trivial software engineering problems in addition to the physics pro...

Descripción completa

Detalles Bibliográficos
Autor principal: Pankratius, Victor
Lenguaje:eng
Publicado: 2012
Materias:
Acceso en línea:http://cds.cern.ch/record/1463354
_version_ 1780925333930770432
author Pankratius, Victor
author_facet Pankratius, Victor
author_sort Pankratius, Victor
collection CERN
description <!--HTML--><p align="justify"> Physics experiments nowadays produce tremendous amounts of data that require sophisticated analyses in order to gain new insights. At such large scale, scientists are facing non-trivial software engineering problems in addition to the physics problems. Ubiquitous multicore processors and GPGPUs have turned almost any computer into a parallel machine and have pushed compute clusters and clouds to become multicore-based and more heterogenous. These developments complicate the exploitation of various types of parallelism within different layers of hardware and software. As a consequence, manual performance tuning is non-intuitive and tedious due to the large search space spanned by numerous inter-related tuning parameters.</p> <p align="justify"> This talk addresses these challenges at CERN and discusses how to leverage multicore parallelization techniques in this context. It presents recent advances in automatic performance tuning to algorithmically find sweet spots with good performance. The talk also presents results from empirical studies conducted with different populations of programmers and programming languages to illustrate which approaches worked well.</p> <h4> About the speaker</h4> <p align="justify"> Victor Pankratius conducts research at MIT in the Computer Science and Artificial Intelligence Laboratory (CSAIL) as a visiting scientist. He is also a Privatdozent at Karlsruhe Institute of Technology where he heads the Multicore Software Engineering investigator group. His research concentrates on how to make parallel programming easier and covers a range of topics including auto-tuning, language design, debugging, software engineering in the cloud, and empirical studies. Victor was also an industry visiting scientist at Intel and Sun Labs/Oracle in the US. He received several awards for his work, including the Intel Leadership Award, for pushing the boundaries at the intersection of software engineering and parallel computing. He serves as the elected chairman of the international Software Engineering for Parallel Systems (SEPARS) working group. Contact him at <a href="http://www.victorpankratius.com" target="_blank">www.victorpankratius.com</a></p>
id cern-1463354
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2012
record_format invenio
spelling cern-14633542022-11-02T22:30:08Zhttp://cds.cern.ch/record/1463354engPankratius, VictorDealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-TuningDealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-TuningComputing Seminar<!--HTML--><p align="justify"> Physics experiments nowadays produce tremendous amounts of data that require sophisticated analyses in order to gain new insights. At such large scale, scientists are facing non-trivial software engineering problems in addition to the physics problems. Ubiquitous multicore processors and GPGPUs have turned almost any computer into a parallel machine and have pushed compute clusters and clouds to become multicore-based and more heterogenous. These developments complicate the exploitation of various types of parallelism within different layers of hardware and software. As a consequence, manual performance tuning is non-intuitive and tedious due to the large search space spanned by numerous inter-related tuning parameters.</p> <p align="justify"> This talk addresses these challenges at CERN and discusses how to leverage multicore parallelization techniques in this context. It presents recent advances in automatic performance tuning to algorithmically find sweet spots with good performance. The talk also presents results from empirical studies conducted with different populations of programmers and programming languages to illustrate which approaches worked well.</p> <h4> About the speaker</h4> <p align="justify"> Victor Pankratius conducts research at MIT in the Computer Science and Artificial Intelligence Laboratory (CSAIL) as a visiting scientist. He is also a Privatdozent at Karlsruhe Institute of Technology where he heads the Multicore Software Engineering investigator group. His research concentrates on how to make parallel programming easier and covers a range of topics including auto-tuning, language design, debugging, software engineering in the cloud, and empirical studies. Victor was also an industry visiting scientist at Intel and Sun Labs/Oracle in the US. He received several awards for his work, including the Intel Leadership Award, for pushing the boundaries at the intersection of software engineering and parallel computing. He serves as the elected chairman of the international Software Engineering for Parallel Systems (SEPARS) working group. Contact him at <a href="http://www.victorpankratius.com" target="_blank">www.victorpankratius.com</a></p> oai:cds.cern.ch:14633542012
spellingShingle Computing Seminar
Pankratius, Victor
Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning
title Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning
title_full Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning
title_fullStr Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning
title_full_unstemmed Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning
title_short Dealing with BIG Data - Exploiting the Potential of Multicore Parallelism and Auto-Tuning
title_sort dealing with big data - exploiting the potential of multicore parallelism and auto-tuning
topic Computing Seminar
url http://cds.cern.ch/record/1463354
work_keys_str_mv AT pankratiusvictor dealingwithbigdataexploitingthepotentialofmulticoreparallelismandautotuning