Eduardo César

As of my incorporation into the Department of Computer Architecture and Operating Systems of the UAB as a postgraduate student, it is possible to divide my scientific-technical career into the following stages:

Simulation of Parallel Applications (1992-99): Focused on the design and development of simulators of parallel applications. This research main objective was the definition of abstractions for parallel programs, based on characterizing tasks and their dependences. Two main abstractions were developed, at first a simpler one, which was easier to parametrize, and, next, a more complex an accurate one. Using these characterizations, several simulation tools were programmed and used in the context of national and European projects. As part of my Master’s thesis, I was involved in the design and development of some of these simulation applications.

National projects: 4, European: 2

International conferences: 3, National: 1, Journal papers: 3

Security in Distributed Systems (2007-12): Focused on the design and development of the FPVA (First Principles Vulnerability Assessment) methodology for the evaluation of vulnerabilities in Grid applications. This methodology clearly defined a set of steps for the assessment of Grid applications vulnerabilities, most of these steps could be automatized or at least supported by specific tools. Jointly with other professors of our group and from the University of Wisconsin, I was involved in the original definition and application of this methodology.

International projects: 2

Master Thesis: 1, Ph.D. Thesis: 1

International conferences: 2, National: 1, Journal papers: 2

Parallel Application Modeling (1999-present): This is my main line of research, aimed at defining high-level performance models for parallel applications. Initially, models were defined for MPI applications with a master-worker and pipeline structure, but later this line has been expanded with the definition of models for memory-intensive OpenMP applications, composed (mix of several structures) applications, applications based on mathematical libraries, distributed data-intensive applications and, finally, applications based on the simulation of agents (ABS) with SPMD structure.

As a result of the work on modeling the performance of ABS parallel systems, we have opened a new line for the definition and implementation of a benchmark for assessing the performance of the parallel simulators generated by well-known platforms, such as FLAME, Repast-HPC or D-Mason. In addition, the knowledge we have gained on this topic has opened new ways of collaboration for optimizing real parallel ABS in the health sciences area (tumor growth and infection spread).

National projects: 12, European: 1

International conferences: 17, National: 4, Journal papers: 11

International Presentations: 4

Parallel Applications Tuning Tools (2010-present): Focused on the design and development of tools for automatic tuning and, in some cases, also dynamic tuning of parallel applications. These tools allow the integration of performance models in the form of external components provided by the analyst. For this reason, this research line is tightly coupled with the Parallel Application Modeling one. The two main tools developed totally or partially by our group are Monitoring Analysis and Tuning Environment-MATE (and its highly scalable evolution ELASTIC) and Periscope Tuning Framework-PTF.

National projects: 2, European: 1

International conferences: 11, Journal papers: 2

Tools: MATE, ELASTIC, PTF

International Presentations: 5

Agent-based modeling and simulation (ABMS) is a class of computational models for simulating the actions and interactions of autonomous agents with the goal of assessing their effects on a system as a whole. Several frameworks for generating parallel ABMS applications have been developed taking advantage of their common characteristics, but there is a lack of a general benchmark for comparing the performance of generated applications. We propose and design a benchmark that takes into consideration the most common characteristics of this type of applications and includes parameters for influencing their relevant performance aspects. We provide an initial implementation of the benchmark for RepastHPC one of the most popular parallel ABMS platforms, and we use it for comparing the applications generated by these platforms.

Agent-based modeling and simulation (ABMS) is a class of computational models for simulating the actions and interactions of autonomous agents with the goal of assessing their effects on a system as a whole. Several frameworks for generating parallel ABMS applications have been developed taking advantage of their common characteristics, but there is a lack of a general benchmark for comparing the performance of generated applications. We propose and design a benchmark that takes into consideration the most common characteristics of this type of applications and includes parameters for influencing their relevant performance aspects. We provide an initial implementation of the benchmark for FLAME one of the most popular parallel ABMS platforms, and we use it for comparing the applications generated by these platforms.

Agent-based modeling and simulation (ABMS) is a class of computational models for simulating the actions and interactions of autonomous agents with the goal of assessing their effects on a system as a whole. Several frameworks for generating parallel ABMS applications have been developed taking advantage of their common characteristics, but there is a lack of a general benchmark for comparing the performance of generated applications. We propose and design a benchmark that takes into consideration the most common characteristics of this type of applications and includes parameters for influencing their relevant performance aspects. We provide an initial implementation of the benchmark for DMASON parallel ABMS platform, and we use it for comparing the applications generated by these platforms.

Contact Tracing Repast HPC agent model

Under development.

This website uses cookies and Google Analytics to help us track user engagement and improve our site. If you'd like to know more information about what data we collect and why, please see our data privacy policy. If you continue to use this site, you consent to our use of cookies.
Accept