A5069833889- Parallel Computing and Optimization Techniques
- Distributed and Parallel Computing Systems
- Interconnection Networks and Systems
- Embedded Systems Design Techniques
- Algorithms and Data Compression
- Advanced Antenna and Metasurface Technologies
- Advanced Data Storage Technologies
- Electromagnetic Simulation and Numerical Methods
- Electromagnetic Compatibility and Measurements
- Machine Learning in Bioinformatics
- Neural Networks Stability and Synchronization
- Electromagnetic Scattering and Analysis
- Simulation Techniques and Applications
- Wireless Body Area Networks
- Advanced Memory and Neural Computing
- Neural Networks and Applications
- Electromagnetic Fields and Biological Effects
- Adaptive optics and wavefront sensing
- Cellular Automata and Applications
- Matrix Theory and Algorithms
- Real-Time Systems Scheduling
- Microwave Imaging and Scattering Analysis
- Metaheuristic Optimization Algorithms Research
- Protein Structure and Dynamics
- Evolutionary Algorithms and Applications
National Agency for New Technologies, Energy and Sustainable Economic Development
2003-2023
AREA Science Park
2008
ENEA Casaccia Research Centre
2002-2008
Tecnologia Energia Ambiente Materiali (Italy)
2005
Sapienza University of Rome
1995-2002
The problem of human exposure to the near field radiobase antennas is a complex and challenging task. A numerical full-wave solution highly recommended because accuracy needed. Unfortunately, computationally heavy. In this paper, parallel finite-difference time-domain tool proposed for addressed so that use supercomputing platforms combined with appropriate phantoms. Results demonstrate good solution. detailed discussion presented on case real antennas, as well some critical issues...
Critical infrastructures (CI) such as telecommunication or the power grids and their dependencies are getting increasingly complex. Understanding these - often indirect -dependencies is a vital precondition for prevention of cross sector cascading failures CI. Simulation an important tool CI dependency analysis, test methods risk reduction, well evaluation past failures. Moreover, interaction simulations with external threat models, e.g., river flood model economic may assist in what-if...
The main goal of Green Flash is to design and build a prototype for Real-Time Controller (RTC) targeting the European Extremely Large Telescope (E-ELT) Adaptive Optics (AO) instrumentation. E-ELT 39m diameter telescope see first light in early 2020s. To this critical component operations, astronomical community facing technical challenges, emerging from combination high data transfer bandwidth, low latency throughput requirements, similar identified barriers on road Exascale. With Flash, we...
Summary form only given. We address the design of pipelined systems able to sustain throughput a given periodic task and, at same time, serve aperiodic requests associated with hard real-time constraints. The proposed method is based on allocation global graph (periodic and tasks), over-dimensioning system devoted process task, deserving unused parts resources management requests. A formal definition such mapping problem, together formalization searching space, space structured in way as...
Abstract We present the architecture of PROSIDIS, a special purpose co-processor designed to search for occurrence substrings similar given ‘template string’ within proteome. Actual tests show speed up figures ranging from 5 50 with respect conventional general-purpose processors. Availability: PROSIDIS configuration file and c code are available at http://www.enea.it/hpcn/php/rosato/
Presents the results of a project aimed at assembling hybrid massively parallel machine, PQE1 prototype, which is devoted to simulation complex physical models. The analysis some existing architectures has revealed that general-purpose machines are largely over-dimensioned and often perform inefficiently in grand-challenge scientific applications. We have thus developed heterogeneous system matches task heterogeneity with architecture heterogeneity: fact, special-purpose architectures, when...
Starting from a mathematical reinterpretation of the classical crossover operator, new type is introduced. The proposed operator gives better performances than 1 point, 2 point or uniform operators. A theoretical investigation behaviour presented. Compared to operators, it allows exploration searching space and results. Some comparative results relative optimization test functions taken literature are given.
The automatic extraction of parallelism from algorithms, and the consequent parallel code generation, is a challenging problem. We present procedure for generation in case algorithms described through SARE (Set Affine Recurrence Equations). Starting original description an N-dimensional iteration space, algorithm converted into (eventually virtual) m-dimensional distributed memory machine (m<N). demonstrate some theorems which are mathematical basis proposed tool. projection technique used...
This paper presents a Knowledge Base System (KBS) as key component of federated simulation framework which allows the investigation (inter)dependencies among Critical Infrastructures (CIs). The KBS supports by using ontological formalism to represent specific CI domains and their dependencies. main advantage proposed consists on abstraction description from technological (i.e. simulation) level: scenario can be described with used experts various domains. approach has been validated...
The Green Flash initiative responds to a critical challenge in the astronomical community. Scaling up real-time control solutions of AO instruments operation specifications modules at core next generation extremely large telescopes is not viable option. main goal this project design and build prototype for an RTC targeting E-ELT first-light instrumentation. We have proposed innovative technical based on emerging technologies High Performance Computing, assessed enabling through prototyping...
In-memory databases that support both online transactional and analytical workloads rely on a process called differential updates to manage writes while maintaining the majority of data in an optimized read-only structure. This, however, necessitates merge write delta partition with main partition, periodically consuming significant amount compute memory resources also causing database downtime during which incoming transaction requests cannot be accepted. In this paper, we present FPGA...
In recent years, the field of Deep Learning has seen many disruptive and impactful advancements. Given increasing complexity deep neural networks, need for efficient hardware accelerators become more pressing to design heterogeneous HPC platforms. The requires a multidisciplinary approach, combining expertise from several areas, spanning computer architecture approximate computing, computational models, machine learning algorithms. Several methodologies tools have been proposed Learning,...
This paper presents a Knowledge Base System (KBS) as key component of federated simulation framework aimed at investigating the dependencies among Critical Infrastructures (CIs). The KBS, based on ontological formalism, represents properties and relations each domain dependency different domains. Some auxiliary data structures, necessary to model interaction simulators CIs, have been defined populated through suitable queries KBS. adoption formalism allowed definition common deal with...
A rigorous theoretical solution of the human exposure to near-field radiobase antennas requires a heavy computational effort. In this paper, we propose full-wave based on implementation finite-difference time-domain method parallel computers, such as APE-Quadrics SIMD machine with 512 processors. The cases several GSM real interacting numerical phantom proposed by Visible Human Project are solved, demonstrating amenability approach and its accuracy for point-evaluation all relevant...
This work provides new perspectives on impact of design effort, consumed resources and abstraction hardware performance in a high-level synthesis flow. We have shown that counter to published literature as well intuition; more effort may not always result better performance. developed kernel simulates Brownian motion, investigated improvement with at various levels. Our results indicate designer should be careful putting particular level. In our case, we achieved best performance/effort...
The paper is concerned with store and forward deadlocks (DL) arising in interprocessor network systems buffered packet switched communications. Algorithms which implement DL free routing use adaptive or nonadaptive modality. Nonadaptive algorithms underuse the interconnection bandwidth because they impose restrictions to paths; are only if certain hypotheses on communication topology occur. In order override these drawbacks, we have implemented an fully exploits connectivity of unrelated its...