Speaker
Describe the added value of the Grid for the scientific/technical activity you (plan to) do on the Grid. This should include the scale of the activity and of the potential user community and the relevance for other scientific or business applications
The systems we are analyzing and evaluating are stochastic; Monte
Carlo
methods are used in order to perform system simulations. These
Monte Carlo
simulations may become extremely computationally demanding as the
physical
dimensions and the loading of the simulated system increase.
Indicatively, a
single simulation may last a day or even more in a standard
desktop computer
and for the convergence of the Monte Carlo to be satisfied,
hundreds of
simulations may be eventually needed. The EGEE infrastructure
provides vast
resources that may be exploited in order to efficiently reduce
execution time.
Use of these resources will enable the simulation and analysis of
more complex
systems that would be infeasible to perform in a desktop computer
or a
conventional small cluster. Monte Carlo methods are widely used
in research
fields (ray-tracing, molecular dynamics, finance, etc) and our
proposed
framework may be used under these contexts with minimal
modifications.
Describe the scientific/technical community and the scientific/technical activity using (planning to use) the EGEE infrastructure. A high-level description is needed (neither a detailed specialist report nor a list of references).
We plan to use the EGEE infrastructure in order to evaluate and
analyze various
techniques and scenarios for the design of mobile communication
systems. Thus
far, we have performed simulations of various scenarios for
Wideband Code
Division Multiple Access (WCDMA) systems. In the future, we
intend to use EGEE
resources for performing further simulations on other mobile
communication
techniques, such as OFDM, smart antennas and multiple-input
multiple-output
(MIMO) techniques.
With a forward look to future evolution, discuss the issues you have encountered (or that you expect) in using the EGEE infrastructure. Wherever possible, point out the experience limitations (both in terms of existing services or missing functionality)
A restraining factor to our approach is that in certain scenarios
a single
simulation may exceed the MaxCPUTime limit mentioned above. In
order to
surpass this problem, we plan to exploit any parallelization
pattern may be
applicable to a single simulation and develop MPI programs for
concurrent
execution at simulation level. This, however, restricts usable
resources to those
of an MPI supporting site, as the support for MPI applications
across different
sites is still missing.
The work presented in this paper has been
funded by the project PENED 2003. The project is co financed 75%
of public expenditure through EC - European Social Fund, 25% of
public expenditure through Greek Ministry of Development -
General Secretariat of Research and Technology and through
private sector, under measure 8.3 of OPERATIONAL PROGRAM
"COMPETITIVENESS" in the 3rd Community Support Program.
Report on the experience (or the proposed activity). It would be very important to mention key services which are essential for the success of your activity on the EGEE infrastructure.
In order to speedup the execution of the simulations, we have
exploited the
parallelizability of Monte Carlo methods and have developed a
task-farming
framework for their parallel execution. The framework is tailored
to the gLite
middleware and implements and manages a task-farming workflow for
parallel
Monte Carlo simulations. Briefly, once the initialization
parameters are given,
the application dispatches several jobs to the grid, assigning a
number of single
simulations to each job. Due to the heterogeneity of the
infrastructure and the
maximum CPU time available for each job, the number of
simulations assigned
to each job is dynamically determined by a simple heuristics that
takes into
consideration execution time and MaxCPUTime and that is
incorporated to the
job itself. Until now, we have executed several scenarios
regarding a WCDMA
system, for various base station configurations, physical
dimensions and system
loading.
