A non-simulated journey has started with Alexander Trujillo

A journey with… whom?

Hi there!, please let me be your guide during this summer… almost forgot! let me introduce myself.

My name is Alexander Trujillo Ochoa, I am a Bachelor of science in physics at Universidad Nacional de Ingeniería, Peru, where I was involved in different projects along with CONIDA (Peruvian Commission for Aerospace Research), one of them was about gamma rays events where I used GEANT4 (a simulation platform based on C++) to simulate cerenkov tanks detectors, another one by simulating plastic scintillator bars in order to collaborate in “Tomography for geophysical objects in Perú” project.

I had an experience at Universidad Nacional Autonoma de Mexico’s Institute of Physics, working at the spectrometer laboratory where I simulated physical events like Rutherford backscattering. Then I got an opportunity to be an intern at General Atomics, USA, there I used computational methods, more specifically, finite element methods, geometry optimization and computer-aided design in a project where I simulated a waveguide system that had safety concerns.

And the last piece of the puzzle is that currently I am studying the Master in High Performance Computing programme in Italy, learning best practices for developing scientific software in projects as multiphysics simulations, so if you see the pattern here is that I love simulations. It mesmerizes me the fact to replicate physical phenomena from real life to computer graphics.

Arriving to Luxembourg

Before continuing, I would like to talk about Luxembourg where I arrived this last thursday, no city I ever been is like this!, could you imagine that here the public transportation is for free? and it is really efficient!. People here normally is polyglot, speaking: French, German, Luxembourgish (Kind of a mix between French and German) and English; so far I had no problem to find my place on this city.

I will work during this summer at the University of Luxembourg, located at Esch-sur-alzette city, for the Computer Science department where I already met my new summer colleagues.

Work ahead

Like I told you some lines ago , I am really fascinated by computer graphic simulations as well as the mathematical and programming framework behind those, but maybe you heard this phrase before… “There is no free lunch!”, we need some computational resources to obtain the precious results.

But it does not end there!, your simulations will not run faster magically if you keep adding more processors to your current resources, it sounds like common sense that parallelizing your work and dividing it into several processes would make you obtain your results faster, but believe or not could make it worse at the end of the day.

This is due to the serial and parallel fraction of your code, in the ideal scenario if your code is embarrassingly parallel most of the work could be easily divided and then the parallel fraction could be higher than 90% , but there will be always a serial part that will ruin our dreams to obtain a ideal linear speedup if we increase the number of processors. This is known as the Amdahl’s law.

My work here at ULux will be related to what I explained before, the project called “A performance comparison and regression for XDEM Multi-Physics Application” has as one of their goals to keep track of the impact on the scalibility for different changes in the code, checking always for the optimal scenario where it is worthy to keep increasing the number of computational resources with the less loss of speed-up.

Now please stay tuned!, and join on this summer journey.

Au revoir!