Pieter de Buck - Week 6 - Duke University

Hi everyone, my name is Pieter and this is my 6th week at Duke.

This week I have started using my own data with my processing scripts. I did this because that meant that I was able to set my own simulation parameters such as beam energy, projectile and target particle and something called the impact parameter. This can be a single number or a certain range of numbers, and it defines the separation of the centers of the two nuclei in femtometers. In the picture below the with of the orange blob is the impact parameter. For my data I used a range from 0.5 to 3.75. I set the input parameters such that they corresponded with the conditions for this RHIC experiment.

I chose to use a fairly large sample size for my purposes, I simulated about 4000 collision events, which took at least 16 hours even on the powerful computer that Duke provided me with. When my research group wants to simulate data for use in scientific papers, they run the simulation in parallel on many different computers, to gain as much precision as possible. I then adapted my scripts to be able to read from multiple files and stitch everything together. This was quite straightforward and allowed me to work on the interpretation of the data very quickly.

Since Dr. Bass left for a trip, I started working with Dr. Nahrgang again. She told me that because of my script she found out that she should not have decreased the precision of the simulation in order to save disk space. In the data it would sometimes occur that two numbers were rounded too early, and thus result in division-by-zero errors. We have started working on new ways to interpret the data, and I will be able to talk about that next week.

I have increased the user-friendliness of my scripts. I added thorough comments, explaining each part of the program. Also the program now reads from an input file, instead of having to define all the parameters inside the program, much like the actual collision software, and allowing for automated running of the script. My research group also created an online repository for each of our scripts, allowing for easy troubleshooting and sharing. Think of it as google docs for programs.

Pieter