Juan Fumero

1.34 Informatics Forum

10 Crichton Street

Edinburgh EH8 9AB

United Kingdom

Read

more

V

Something about me

I am studing my third year of the PhD in the CArD - Compiler and Architecture Design Group in the School of Informatics at The University of Edinburgh under the supervision of Christophe Dubach. I am a member of the ICSA - Institute for Computing Systems Architecture. My PhD is funded by Oracle Labs.

My work is focused on runtimes systems for heterogeneous computing. I am using Graal, a new JIT compiler developed by Oracle Labs to apply compiler transformantions and code generation for OpenCL.

Research Interests

My research is focused con Compilers and Runtime Code Generation for Heterogeneous Computing. I am interested in:

 

  • Compiler techniques.
  • Software design.
  • Virtual Machines.
  • Data management for heterogeneous systems.
  • Heterogeneous and parallel computing.

Internships

  • June - Nov 2015: Oracle Labs intern (Linz, Austria) : FastR and Apache Flink. Compiler support for distributed data and batch processing.
  • July - September 2013: CERN Openlab 2013 (Geneva, Switzerland): Vectorization with Intel Haswell Microarchitecture.
  • September - Nov 2012: Edinburgh Parallel Computing Center (EPCC, Edinburgh): HPC-Europa2 scholarship. accULL compiler andr untime for OpenACC (multi-device).
  • July 2012 - June 2013: Research Institute for Renowable Energy (Spain) : system administration for HPC service.
  • September 2010 - July 2012: SAII (La Laguna University - Spain): system administration for HPC service.

Teaching

  • Tutor: Object Oriented Programming (Java course), The University of Edinburgh. Code: INF1-OP 2016
  • Tutor: Object Oriented Programming (Java course), The University of Edinburgh. Code: INF1-OP 2015

Invited Talks

  • 18th May 2016, University of Edinburgh. Edinburgh R Users Group FastR-Flink
  • 18th January 2016, TU Berlin (Germany). FastR-Flink. A compiler based approach for distributed computing in R.

Research Groups

  • Organizer PhD student talks at Informatics Forum. Monthly technical talks for PhD students in Computer Science.

Publications

2017

  • Juan Jose Fumero, Michel Steuwer, Lukas Stadler and Christophe Dubach. Just-In-Time GPU Compilation for Interpreted Languages with Partial Evaluation Accepted in VEE'17
  • Juan Jose Fumero, Michel Steuwer, Lukas Stadler and Christophe Dubach. OpenCL JIT Compilation for Dynamic Programming Languages Accepted in MoreVMs 2017

2015

  • Juan Jose Fumero, Toomas Remmelg, Michel Steuwer and Christophe Dubach. Runtime Code Generation and Data Management for Heterogenous Computing in Java PPPJ'15.  |
  • Juan Jose Fumero, Michel Steuwer and Christophe Dubach. Enabling Heterogeneous Computing in Java ECOOP Truffle Workshop 2015.  |
  • Juan Jose Fumero, Michel Steuwer and Christophe Dubach. API and Runtime System for Heterogeneous Composable in Java [POSTER] at HiPEAC 2015  |

2014

  • Juan Jose Fumero, Michel Steuwer and Christophe Dubach. A Composable Array Function Interface for Heterogeneous Computing in Java: PLDI ARRAY 14.  |

2013

  • Juan Jose Fumero, Andrzej Nowak. Vectorization with Haswell and CilkPlus. Report at CERN openlab 2013. August 2013. 
  • R. Reyes, Juan J. Fumero, J. Lucas and Francisco de Sande. Programming for GPUs:  the Directive-Based Approach. 8th International Conference on P2P, Parallel, Grid, Cloud and Internet Computing.
  • R. Reyes, Ivan Lopez, Juan J. Fumero and Francisco de Sande. State of the Art in OpenACC Implementations. Journal of Supercomputing. Year: 2013. DOI: 10.1007/s11227-012-0853-z.

2012

  • R. Reyes, Ivan Lopez, Juan J. Fumero and Francisco de Sande. An Early Evaluation of the OpenACC Standard. 12th International Conference Computational and Mathematical Methods in Science and Engineering.
  • R. Reyes, I. Lopez, J. J. Fumero and  F. de Sande. accULL: An OpenACC implementation with CUDA and OpenCL support. International European Conference on Parallel and Distributed Computing (EuroPar 2012).
  • R. Reyes, I. Lopez, J. J. Fumero and  F. de Sande. accULL: an user-directed Approach to Heterogeneous Programming. The 10th IEEE International Symposium on Parallel and Distributed Processing with Applications.
  • Ruyman Reyes, Ivan Lopez, Juan J. Fumero and Francisco de Sande. Directive-based Programming for GPUs: A Comparative Study. HPCC-2012.

External Links

2017 | Juan Fumero

Designed by Pablo Anibal del Rosario: webpage