The thematic Iberian-American Network for High Performance Computing (RICAP) is presented to provide the region with a strategic infrastructure in the field of high performance computing from an advanced architecture that includes both High Performance Computing (HPC) and High Throughput Computing (HTC).
RICAP brings together several computer centers, various use case providers in different scientific and technological fields, one of the world's largest supercomputing companies and an experimental Latin American consortium in the field of physics.
This Thematic Network will also develop different software tools aimed at facilitating access and computational efficiency in this hardware infrastructure and will encourage its free use through different dissemination and outreach actions that are expected to attract users of the different scientific and industrial areas. With all of this, a high computing power will be made available to the community, which is a real alternative to proprietary services located outside the region.
Specific objectives include: effective interconnection of high performance open services from RICAP clusters (both supercomputing and cloud access); the implementation and subsequent promotion of software-based solutions for the access and exploitation of this network; the design and development of open source tools that will improve the computational efficiency of the infrastructure in an unattended and dynamic way, especially in an environment such as the cloud; the promotion of the transfer of knowledge and the impact of RICAP through the provision of tutorials and seminars for administrators and end users with the latest technologies in the fields of HPC and HTC; and, the collaboration with other national and regional initiatives (RedCLARA, H2020 and others).
With the achievement of these objectives, it will be possible to carry out new activities by different Latin American groups for whom it was previously very difficult to perform their work in data simulation or analytics because they lacked enough computing power.
From January 2017, these are the milestones that have been identified by RICAP (the yearly fulfilment of these KPIs can be found in italics):
|Call for accessing HPC resources||1||1||2||2||6|
|CPU hours used (x1000)||200||200||400||400||1200|
|208.25||2,227 CPU & 10,000 GPU||4.879 CPU & 17.250 GPU||4.460 CPU & 20.000 GPU||10.549 CPU & 47.250 GPU|
|External publications and presentations acknowledging RICAP||5||5||10||10||30|
|Groups accessing the infrastructure||5||5||10||10||30|
|2 schools||5 schools||5 schools||1 school||13 schools|
Also, the following deriverables will be published:
|D1||Website of the thematic network||M2|
|D2||Definition and publication of the first call for accessing HPC resources||M3|
|D3||Cloud infrastructure operational||M6|
|D4||RICAP solutions for an efficient use of computational infrastructures||M12|
|D5||First year training and dissemination actions||M12|
|D6||First year results of the use of the HPC & HTC infrastructure||M13|
|D7||Publication of the second call for accessing HPC resources||M15|
|D8||Analysis of the RICAP solutions for an efficient use of computational infrastructures||M17|
|D9||RICAP solutions for accelerators||M24|
|D10||Second year training and dissemination actions||M24|
|D11||Second year results of the use of the HPC & HTC infrastructure||M24|
|D12||Publication of the third call for accessing HPC resources||M25|
|D13||Analysis of the RICAP solutions for accelerators||M28|
|D14||Publication of the fourth call for accessing HPC resources||M31|
|D15||RICAP solutions for fault tolerance in the cloud||M36|
|D16||Third year training and dissemination actions||M36|
|D17||Third year results of the use of the HPC & HTC infrastructure||M37|
|D18||Publication of the fifth call for accessing HPC resources||M37|
|D19||Analysis of the RICAP solutions for fault tolerance in the cloud||M40|
|D20||Publication of the sixth call for accessing HPC resources||M43|
|D21||RICAP solutions for containers||M48|
|D22||Fourth year results of the use of the HPC & HTC infrastructure||M48|
|D23||Fourth year training and dissemination actions||M48|
(M2 stands for the second month of the Project, i.e. February 2017 and so on).
Project Structure and Committees
RICAP is managed through an Executive Committee that is formed by a representative of each member group.
The Committee members are:
CIEMAT – Rafael Mayo García (Coordinador)
BSC-CNS – Ulises Cortés García
CEDIA - Claudio Chacón Arévalo
CeNAT-UCR – Álvaro de la Ossa Osegueda
CINVESTAV – Isidoro Gitler Goldwain
CSC-CONICET – Esteban Mocksos and Alejandro Soba
CUDI - Israel Novelo Zel
Fujitsu TS – Adriano Galiano Díez
HPC-Cuba – Héctor Cruz Enríquez
LAGO - Dennis Cazar Ramírez
NLHPC - Ginés Guerrero
UDELAR - Sergio Nesmachnow
UFRGS - Philippe O.A. Navaux
UIS – Carlos Jaime Barrios Hernández
Uniandes - Harold Castro Barrera
This Executive Committee will appoint experts to evaluate the proposals submitted to the calls for accessing the HPC resources, who will be made public jointly with the call resolution.
RICAP counts on Salma Jalife Villalón (Mexico) and Andrea Morales Rodríguez as CYTED supervisors.
How to acknowledge RICAP
Whenever you will publish works in articles, conferences or any other media in which you have made use of RICAP resources, add this sentence into the Acknowledgements Section: "This work has been partially carried out with resources provided by the CYTED cofunded Thematic Network RICAP (517RT0529)."