Intel® Parallel Computing Center at São Paulo State University (Universidade Estadual Paulista, UNESP)

Principal Investigators:

Sérgio F. Novaes

Sérgio F. Novaes is Professor of Physics at the São Paulo State University (UNESP) and Scientific Director of the Center for Scientific Computing (NCC/UNESP). He carries out research in High Energy Physics: he was member of the DZero Collaboration from Fermilab, and is currently part the Compact Muon Solenoid (CMS) Collaboration from CERN. He is the group leader of the São Paulo Research and Analysis Center (SPRACE) and coordinated the deployment of GridUNESP, the first Campus Grid in Latin America. He is the Brazilian representative at the Particles and Fields Commission (C11) from the International Union of Pure and Applied Physics (IUPAP).


Geant (GEometry ANd Tracking) is the toolkit of choice for the simulation of particle interaction with matter. Its development started almost twenty years ago and has evolved since then with the effort of an international collaboration of physicists and computer scientists from many institutions. Developers interact constantly with users in a combined effort to validate the results for application in high energy physics experiments, space projects and medical studies. Geant incorporates the knowledge from modern particle and nuclear physics and it is designed to model all the elements associated with particle detector simulation: the geometry of the system, the electromagnetic fields inside the materials, the physics processes governing particle interactions with matter, the response of sensitive detector components, the visualization of the detector and particle trajectories, and the capture and analysis of simulation data at different levels of details and refinements. The general-purpose nature of Geant for analyzing interactions of particle with matter goes beyond particle physics detectors. It has been used for instance in space applications (study of the effect of space radiation on hardware components and humans), medical applications (used for dosimetry, emulating the real radiotherapy treatment), and radiation-hard electronics (modeling the radiation damage in microelectronics). Geant4, the current version of the Geant series of software toolkits, was the first to use object oriented programming techniques.

The Unesp IPCC will be mainly involved in R&D efforts to adapt HEP software tools, including Geant, in order to explore the modern computing architectures that support multi-threading and other parallel processing techniques to make data processing more cost-effective. Geant requires long calculation times and it is ideal for compute-bound workloads that is suited for execution on Intel Xeon Phi coprocessors. The plan for code performance improvements at Unesp IPCC is focused on testing vector-coprocessor prototypes in hybrid parallel computing systems and analyzing the performance of the next generation of Intel Xeon Phi coprocessor (Knights Landing), evaluating the redesign efforts eventually necessary for adopting this new technology. The activities will be closely related to the development of Geant-V, the next generation of the Geant simulation engine, which will include massive parallelism natively. The multi-threaded version of Geant4 and Geant-V will ultimately converge to a common code

Related websites:

For more complete information about compiler optimizations, see our Optimization Notice.