Servicio de Información Comunitario sobre Investigación y Desarrollo - CORDIS

Final Activity and Management Report Summary - POT (Precise Online Tracking for high energy physics event selection at CDF)

During the Marie Curie fellowship I focused on studying and improving the trigger system for High Energy Physics Experiments (HEP). The trigger system is crucial to the success of hadron collider experiments. For example, the Tevatron at Fermilab produces about 10 million proton-antiproton collisions per second. A maximum of 100 can be stored for data analysis. It is the job of the trigger to select the most interesting 0.001% in real time, reconstructing and filtering the event in a timing period of the order of few microseconds.

As a goal of my research I wanted to pursue a trigger strategy which minimises the difference between the online and the offline performance in event reconstruction: this is key to cope with the increasing luminosity (more complicated events) at the Tevatron and to increase the discovery potential of the CDF experiment (Outgoing Phase) and it will be even more crucial for LHC experiments at CERN (Ingoing Phase). A massive computing power is required to minimise the online execution time of complex offline algorithms. This can be obtained using new powerful technologies: commercial CPUs and sophisticated programmable devices as Field Programmable Gate Arrays (FPGA) and ASICs. This strategy was applied to the most recent upgrades of CDF.

As hardware expert, I took the following responsibilities: 1) Level-2 Calorimeter Upgrade, as technical coordinator; 2) Level-1 Calorimeter Upgrade, as project leader; 3) GigaFitter Project, as consultant; 4) Silicon Vertex Tracker system coordinator and proponent of a SVT Upgrad; 5) Level-2 Outer Tracker system Upgrade, XFT project, as consultant.

All the previous Upgrades have been completed successfully and the new systems are now the official CDF trigger systems with advantages both in terms of the operational functionality (the experiment runs all the day long trying to optimise the data acquisition) and physics results). I also took responsibility first as 6) Hardware Trigger Group Leader, then as 7) Trigger Dataset Co-Head; this last role implied interfacing with physics groups and planning the trigger strategies of the experiment, weekly reports to the CDF collaboration and to the experiment staff. Due to the successfully trigger upgrades at CDF it's natural to extend such a strategy to Large Hadron Collider experiments at CERN. However due to much higher occupancy and event rates and the fact that the detectors are much more complex, there is an enormous technological challenge.

I'm currently the technical coordinator of an Italian R&D activity to develop a new generation of Content Addressable Memory (CAM) device optimized to perform pattern recognition of particles tracks at ATLAS experiment at CERN. Moreover new promising applications of the device even outside HEP have been studied and presented to important recent conferences. I've also recently started a new collaboration with AuroraScience project.

AuroraScience is an European researcher project at the crossroads of computational sciences and computer architecture with the goal to define a next generation high performance computing for scientific and technological applications. Physics has always been driving High Performance Computing (HPC) innovation: the APE supercomputing family originated by INFN has been for more than 20 years one of the most successful European HPC initiatives in computational particle physics (my main activity during my PhD studies). AuroraScience continues this experience extending to new scientific area and involving new advanced technologies. My goal is to study and to develop a new and outstanding CPU-CPU data transfer for two main classes of physics applications: Computational Fluid Dynamics (CFD) and Lattice Quantum ChromoDynamics (LQCD) simulations.

Reported by

Síganos en: RSS Facebook Twitter YouTube Gestionado por la Oficina de Publicaciones de la UE Arriba