Day 1, SC15

We started the day by sitting in on a panel session on research in HPC with academics from across the USA. Whilst post graduate studies and advanced degrees are a little over the horizon for our guys, the overall message of love what you do and have a passion for your work was not lost on us.

The students were also encouraged to have "blue-sky" ideas and to not be "code-slaves". Creativity and inspiration count for something even in the product-driven world of HPC. One professor made the statement that if you present your work to someone and they tell you it's stupid, pursue it. It may be the great idea for which you become known.

The concept that you should focus on one or two things early in your HPC career so that you gain great depth of knowledge then, once recognised in the field, quickly broaden your focus because someone in HPC with a broad understanding of all Science domains is invaluable.

The second session we attended explained the different aspects of the SC conferences and how they related to students. There was some information we found quite useful in fleshing out the rest of the conference sessions we shall attend.

We then broke for lunch and had a walk across the Colorado in search of cowboy boots for me. Unfortunately we were unsuccessful.

After our walk we settled in to another student session where Alan Sussman gave a quick overview of parallel programming - hardware, software, techniques and data partitioning. Although a whirlwind tour of HPC, the guys found it quite easy to understand and it engendered quite some discussion on parallel techniques, data partitioning and how we were going to apply this when we build our own parallel programming platform.

The last session was a plenary session with Intel Fellow Diane Bryant who discussed the next step in supercomputing - Exascale. This is some three orders of magnitude from where we are at the moment and is predicted to happen early in the 2020s. There are many challenges we need to face before this becomes a reality. Diane's talk was engaging and interesting and included a number of Intel Fellows who joined her on the stage. One of the highlights was a photograph montage of Gordon Moores, co-founder of Intel and the author of Moore's Law, a principle and perhaps self-fulfilling prophecy he developed in the 60s that relates to the number of transistors we can fit onto an Integrated Circuit (IC or chip) doubling every 2 years. Moore's Law is a foundational principle in computing and has held true for the past half-century.