My SoHPC Journey: The Beginning

Hello! My name is Iakov, I am from Russia, but I study in the UK. A couple weeks prior to me writing this self-introduction, I finished my three-year-long Geophysics degree. During my studies, I’ve had a chance to acquire some computational skills, apart from all things “Geo”. Having a more “computing inclined” mindset, I naturally began the search for summer internships which would allow getting some “hands on” programming experience.

After the typical nerve-wracking process of sending out applications, I was delighted to have been selected for the project 2124 in Summer of HPC by PRACE. That really was great because this programme was an amazing opportunity to spend a summer learning a more about computing and testing scientific ideas on a HPC system as part of a research project (you don’t get to do that every day!)

Thoughts on the project

It is possible for the name of the project I will be working on to sound a bit terrifying: “Hybrid AI Enhanced Monte Carlo Methods for Matrix Computation on Advanced Architectures”. However, I was immediately attracted to it because it combines Machine Learning, Monte Carlo methods and inversion problems, all of which I am interested in. At university, I took a class, specifically dedicated to inversion problems in Geosciences and the basics of their numerical implementations. Such problems are very important in the study of the Earth where modelling a medium with large number of physical parameters is one of the key approaches to research. A traditional example is inverting the data recorded by seismometers (instruments sensitive to ground shaking from earthquakes or other seismic events) to obtain velocity structure of the Earth. That, in turn, can tell us a lot about what goes on below our feet. Do not be worried, that is all for the digression into the field of Geophysics for now.

Unlike what has been described above, this SoHPC project is theory oriented. There is a lot to do without even dealing with applications. This summer is definitely going to be challenging for me, but that’s a good thing. The best way to get better at something or to learn new concepts is simply to get out there and start doing! I expect to invest time into chasing up different bits of theory and figuring out how things work.

First impressions

It remains to be seen what my work routine for the rest of the summer will be, but for now, we are past the training week! It was great to learn about such powerful tools as OpenMP and MPI, along with Cython, Numba and more. I want to say special thanks to the instructors from Irish Centre for High-End Computing (ICHEC) who organised the whole event and presented the material in a superb way. So far, I have only had one meeting with my project mentor, co-mentor and the rest of the team. I am looking forward to progressing with my introductory reading and seeing what comes next.

It has not been long since the start, but it is already clear that this journey is an exciting one. Stay tuned for updates on how the project is coming together and an explanation of what exactly I will be doing.

And to my fellow SoHPC participants: best of luck and enjoy!