An overview of my project

The quarks that are bound together to comprise a proton

Now I have finished my time in Germany, I think it’s time I told you all a little about what I was actually doing there.  My project was entitled “Making Quarks Phli Further” and was basically on quantum chromodynamics (QCD). QCD is the theory of the strong interaction, one of the four fundamental forces of nature. It describes the interaction between quarks and gluons which make up hadrons such as protons and neutrons. Now, what is a quark? These are the fundamental building blocks which come together to form composite particles such as protons and neutrons. ie. a proton is formed of two up quarks and one down quark. These interact via gluons (the purple lines in the diagram)

In normal low temperature/density conditions quarks and gluons are permanently confined in the hadrons. This means that a single quark cannot be isolated by a microscopic distance.  The equations that govern QCD are very complicated and therefore not directly solvable, to solve them an approximation must be used. The technique most often used to solve these complicated problems is Lattice Quantum Chromodynamics (LQCD), which regularises QCD by introducing a space-time lattice.

However, even this has its downside, it requires extremely large amounts of computational power to solve even very small systems. `the largest source of slowdown occurs because it takes time for the computer to access the memory. This depends on where the data is stored. Data can be stored in main memory or in cache memory.  The cache is a smaller, faster memory which stores copies of the data from frequently used main memory locations. This is used by the CPU of a computer to reduce the average time to access data from the main memory. There are usually three different levels of cache memory each with different memory access times.  Ideally you want all the memory that you require to be stored in the cache so the computer never has to access the main memory.

The access time for the different cache levels

The project focussed on benchmarking LQCD code, to find places where the memory was not being efficiently managed/stored. This will hopefully speed up the programme and mean that systems can be solved with larger numbers of quarks and gulons.

Now, I am back in England, I have had time to think about what an amazing experience ASoHPC was. Whilst being a great opportunity to learn a lot about a subject area that previously I knew very very little about it was also great to meet some amazing people both in Barcelona and in Germany, they really made this summer the experience that it was. Therefore, I would like to thank everyone who organised a SoHPC, the Barcelona supercomputing centre for hosting us during the training week (especially for all the much-needed coffee), everyone at Jülich who helped and supported me through my project especially my supervisor Stefan for his patience when I didn’t understand anything (and for the amazing barbeque).