Error rendering component

Guest lecture: "Research Experience. Energy-Efficient Computing Vision" - Dr. Pedro Valero-Lara, Computational Technology Unit, BCAM

Dr. Pedro Valero-Lara, Computational Technology Unit, Basque Center for Applied Mathematics (BCAM), vil holde en gjesteforelesning fredag 18. september 2015 kl 10:15. Sted: Tabletten (F1.101), Farmasibygget, Campus Tromsø.
Fuglesteg, Jan
Published: 14.09.15 00:00 Updated: 14.09.15 09:15

Faglig kontaktpersoner er førsteamanuensis Phuong H. Ha

Kart over Campus: ( https://uit.no/campustromso )

Alle interesserte er hjertelig velkomne.

Påmelding er ikke nødvendig.

Vel møtt!

--------------------------------------------------------------------------------------------------------------------------------------------

Date:
Friday September 18th 2015, 10:15 
Place: Tabeletten (F1.101), Farmasibygget, Campus Tromsø 
Title: " Research Experience. Energy-Efficient Computing Vision
Lecturer: Dr. Pedro Valero-Lara, Computational Technology Unit, Basque Center for Applied Mathematics (BCAM)

Abstract:
This talk covers the major goals achieved along many years working in multiple fields and computational platforms, projects where I have participated, frameworks developed, programming environments and computational platforms used, etc. In particular, the presentation includes several applications such as, Image Processing (Brain and Satellite Images), Metric Spaces Methods and Longest Common Subsequence Implementations for similarity search, Poisson Solvers and Computational Fluid Dynamics solvers. All these applications are evaluated over heterogeneous platforms, so this talk covers a large spectrum of applications, and how theirs particular features can be efficiently implemented over such computing platforms. In some of the applications is also included a power consumption study. Besides, it is included a performance study about one of the most popular Partitioned Global Address Space (PGAS) programing model, that is Unified Parallel C (UPC), against a hybrid MPI-OpenMP model for Lattice-Boltzmann Fluid Simulations. After that, it is presented some ideas about Energy-Efficient Computing in Many-core architectures. In particular, it is analyzed the advantages of the current GPUs in terms of power-consumption, and the upcoming advances. Finally, at the end the talk, I will present the current and upcoming projects in which I am involved today. 

Short bio:
Pedro Valero-Lara received his B.S. and M.Sc. degrees in Computer Engineer from the University of Castilla-La Mancha (Spain, 2009 and 2010) and his Ph.D. degree in Computer Science from the Univ. Complutense of Madrid (Spain, 2015). His Ph.D. comprises heterogeneous computing and efficient implementations of large-scale scientific problems. He is researcher at Computational Technology Unit, BCAM. Although his experience includes several fields, such as Image Processing or Data Retrieval, his main research line is High Performance Computing applied to Computational Fluid Dynamics. His work covers the use of different parallel systems, such as NVIDIA GPUs, Intel MIC, multicore, Cray, among others. 

He is the main developer of the LBM-HPC software framework for fluid simulations based on Lattice-Boltzmann Method. He has authored over 30 publications, including 4 journal citation report (JCR) papers and more than 20 peer-reviewed international conference papers. 
  

We can recommend