SDC Presentation – FermiLab Software Development

On Sunday, October 5th, I attended the Software Development Community’s (SDC) presentation on software development at FermiLab, presented by Dr. James Amundson. It was a fascinating look into the world of high performance computing and clustering. It’s my first time attending SDC which there was a good sized crowd and everyone seemed genuinely interested in the topic at hand and there were many good questions. I wish I wrote some of them down.
In any case, there were numerous interesting facts about software development at FermiLab so I won’t rehash all the details. There was lots of discussion on highly parallel processing, clustering and large data sets – apparently this is par for the course for them. They use C++ and Python with lots of unit testing. One thing I found extremely interesting was the amount of data that was collected per “event” within a simulation (shown on the left): 1 – 25MB of data @ 40MHz = 1 Petabyte/s of raw data. That’s insane. I thought terabytes were large. It puts things into perspective when doing business applications on daily and then realizing how vastly different scientific research resource needs are. I thoroughly enjoyed the presentation. Dr. Amundson gave. It was a very interesting insight into the scientific research software development field. You can see more from the SDC site on Meetup.com here. Hopefully I’ll