We are excited to announce the first GPU Mini-Hackathon in Costa Rica!
General Purpose Graphic Processing Units (GPUs) offer high performance gains for a wide range of scientific applications. Modern accelerator architectures provide highly parallel capabilities which, the application programmer must learn to exploit. These hackathons are intended to help new and experienced GPU programmers overcome the challenge of exposing and expressing enough parallelism in their applications to speedup their execution.
The goal of this event is that current or prospective GPU programmers learn how to use these accelerators to design and optimize (potentially) scalable scientific applications. Throughout this 3-day hands-on hackathon, experienced mentors will help attendees leave with extensive examples of applications running on GPUs and a clear road map of how to jump-start their own applications to run on GPUs.
Previous programming experience with OpenACC or CUDA is not required. However, exposure to traditional HPC programming tools (MPI or OpenMP) is desirable. Applicants with a target application to port or optimize on a GPU are highly considered. Work will be done in teams of 2-3 developers.
|DAY & TIME||Monday 21st||Tuesday 22nd
|8:00 – 10:00||OpenACC Lectures||Mini-App Challenge||Bring your own code|
|10:00 – 10:30||Morning Break|
|10:30 – 12:00||OpenACC Lectures||Mini-App Challenge||Bring your own code|
|12:00 – 1:00||Lunch Break|
|1:00 – 3:00||Mini-App Challenge||Bring your own code||OpenMP 4.5/ Kokkos|
|3:00 – 3:30||Afternoon Break|
|3:30 – 5:00||Mini-App Challenge||Bring your own code||OpenMP 4.5/ Kokkos|
Biography: Oscar Hernandez received a Phd in Computer Science from the University of Houston. He is a staff member of the Computer Science Research (CSR) Group, which supports the Programming Environment and Tools for the Oak Ridge Leadership Computing Facility (OLCF). He has experience in creating and driving many de-facto standards such as OpenACC, OpenSHMEM, UCX and participates in the OpenMP community. At ORNL he works closely with application teams including the CAAR and INCITE efforts at OLCF to address their programming model and tools need. He is currently working on the programming environment for Summit, the world’s fastest supercomputer (as of Nov 2018), and works very closely with the vendors on their plans for next-generation programming models. He has worked on many projects funded by DOE, DoD, NSF, and Industrial Partners in the Oil & Gas industry.
Jose Manuel Monsalve Diaz
Biography: Jose Manuel Monsalve Diaz obtained his bachelor degree in Electrical Engineer from the Pontificia Universidad Javeriana in Bogotá in 2013. During his bachelor he worked as a research assistant for KM-RoBoTa s.a.s. where he helped develop a simulator for a Modular Snake-inspired Robot for the LoLa Open Platform. After graduation, he moved to the University of Delaware in 2013 where he is currently pursuing his doctoral and master degree on the areas of Parallel Computing and Parallel programming. Throughout these years he has worked as a research assistant of the CAPSL research group for Prof Guang. R. Gao, and the CRPL research group for Prof. Sunita Chandrasekaran. His area of interests are parallel programming models covering conventional programming models such as OpenMP and OpenACC for CPU and heterogenous systems, as well as unconventional Data-flow based programming models and computer architectures such as the Codelet Model developed at the University of Delaware.
Biography: Jeff Layton is an HPC Advocate at NVIDIA focusing on OpenACC and Hackathons. All of his degrees are in Aeronautical and Astronautical Engineering. In his 30+ years of experience he has been a professor, engineer, cluster builder, cluster user and cluster admin, code writer, system architect/engineer, manager, and benchmark/IO engineer. He’s a big proponent of GPUs, MPI, OpenACC, CUDA, Python, Deep Learning, and Fortran and loves helping people solve problems and help people use new technologies to improve their applications.
Pedro Mário Cruz e Silva
Biography: Pedro Mário Cruz e Silva did his BSc (1995), and MSc (1998) at Federal University of Pernambuco (UFPE), he also did his DSc in 2004 at PUC-Rio. Worked for 15 years as Manager of Computational Geophysics Group at PUC-Rio, during this period was responsible for several Software Development and R&D projects for Geophysics with strong focus on innovation. He also finished an MBA in 2015 at Getúlio Vargas Foundation (FGV/RJ). Is member of the main board of The Brazilian Geophysical Society (SBGf). Currently is the Solution Architect Manager at NVIDIA responsible for all technologies in the Latin America Region.
João Paulo Navarro
Biography: João Paulo Navarro is a Computer Scientist and M.Sc. in Computational Modeling by UFJF, Brazil. For years he was a member of the Computer Graphics Group of UFJF, involved in researches in the areas of visual computing, physical simulation and high-performance computing. Experienced software engineer, had developed for many years systems at PUC-Rio, focusing on visual computing and machine learning applications for Petrobras Oil & Gas company. He is author of publications at international conferences in machine learning and computer graphics, which is his specialties. Currently, works as Solutions Architect at NVIDIA, mainly focused on GPU computing at scale, code optimization, machine learning and deep learning in high-performance environments.
Carlos J. Barrios
Biography: Carlos J. Barrios is an associated professor of the Universidad Industrial de Santander (Colombia) and director of the High Performance and Scientific Computing Center in the same university. Doctor in Computer Science (Université de Nice-Sophia Antipolis, France), Master in Applied Mathematics and Informatics (Université de Grenoble, France) and Systems Engineer (Universidad Industrial de Santander, Colombia), he has experience in HPC and Scientific Computing from several years. He is co-chair and co-founder of different outreach and scientific events as the Supercomputing and Distributed Systems Camping School (www.sccamp.org) and the Latin American High Performance Computing Conference (www.ccarla.org). Today, Carlos J. is the general Chair of the Advanced Computing System for Latin America and Caribbean (SCALAC) and also he is NVIDIA Deep Learning Institute University Instructor (More information in: https://www.linkedin.com/in/carlosjaimebh/)
Participation is free of charge. Lunches and accesses to computational infrastructure throughout the event are offered by the participating organizations.
The maximum quota is 60 participants.
The following form has to be fully filled before January 6, 2019. Accepted participants will be notified via email on January 7, 2019.
Start of the application process: November 26, 2018
Closure of the application process: January 6, 2019
Notification of acceptance/rejection: January 7, 2019
Travel and Hotel Information
Transportation: Taxis or Uber from the airport to the hotels are available. An uber ride from the airport to one of the listed hotels costs around 10-12 dollars.
An uber ride from the hotels to the event costs around 3-4 dollars.
|Hotel||Webpage||Distance to CeNAT||Phone Number|
|Palma Real Hotel & Casino||https://www.hotelpalmareal.com/||3.5 km||(+506) 2290 5060|
|Hotel Parque del Lago||http://www.parquedellago.com/hotel.html||4.1 km||(+506) 2547-2000|
Hilton Garden Inn San Jose La Sabana
Crowne Plaza San Jose Corobici
Some more economic options are available:
|Hotel||Webpage||Distance to CeNAT||Phone Number|
|Isla Verde Hotel & Restaurante||http://www.hotelislaverdecostarica.es/||1.0 km||(+506) 2232-5203|
|The Rock Bed & Breakfast||https://www.thebnbtherock.com/||100 m||
|Casa Lima B&B||http://www.casalimacr.com/||800 m||
(+506) 2231 0064
In accordance with the General Act on Migration and Inmigration, foreign persons who intend to enter in Costa Rica must bear:
- A passport or valid travel document
- A visa if it is required
- Proof of economic sufficiency in a minimum of USD$100.00 for each month of legal permanence in the country
- A bill, ticket or voucher for the continuation of their trip
- Not havin an impediment to enter in the Costa Rican territory
People from Colombia, Dominican Republic and Ecuator require a visa to enter Costa Rica. A full list of countries with visa requirements and further information can be found here.
However, nationals of those countries with a visa requirement that may fit in any of the waivers or conditions for entry which are stipulated as follows, may be dispensed from the visa procedures to Costa Rica:
- Nationals of those countries with a consular or restricted visa requirements who possess a multiple entry visa(tourism visa, crew member visa of business visa) to the United States of America(exclusively a B1-B2 type visa, a D type visa or a C1/D multiple entry visa) or Canada, may be dispensed from the consular visa for the purpose of entering into Costa Rica.
- Nationals of those countries with a consular or restricted visa requirement with a legal permanence that allows for multiple entries and a minimum validity of six months in the United States of America, Canada and the countries of the European Union, may be dispensed from the consular visa for the purpose of entering into Costa Rica.