Here's part three of our interview with Professor Thomas Sterling.
Sterling is director of the Center for Research in Extreme Scale Technologies (CREST) at Indiana University and has been integral to the development of high-end computing, most notably his groundbreaking work with the Beowulf cluster.
You can find part one here, where Sterling discusses how HPC has evolved to its present state. The second installment, in which Sterling forecasts the future course of HPC, can be read here. And look here for our preview of ISC 2017.
This week, Professor Sterling reflects on being part of the ISC family and shares some anecdotes from over the years.
Now in its 32nd year, how would you describe the value of ISC to the HPC community?
ISC has provided the single most important venue for realizing a balanced international HPC community. It complements the Supercomputing series (SC) in the US by delivering an emphasis and culture as viewed from outside the US.
Over its many years, ISC has also experimented with different forms of engagement, and the impact has been very visible.
For example, it is the founder of ISC, the late Hans Meuer, who, with colleagues Jack Dongarra, Horst Simon, and Erich Strohmaier, adopted Jack’s Linpack benchmark to create the Top500 List that has been used by the HPC community to rank supercomputers since 1993.
ISC has emphasized invited talks of key leaders and contributors of timely accomplishments. Through the years, industry vendors in hardware and software have engaged in tough question-and-answer sessions and direct narratives of industry goals and methods.
A high level of attention is also given to major program thrusts of many national research centers and labs throughout Europe, Asia, and the southern hemisphere.
Every year for the last 14, the final keynote address is dedicated to a yearly retrospective, a kind of op-ed piece about trends and accomplishments of the last year. An annual student cluster competition is highlighted, featuring candidate undergraduate teams from all over the world.
In short, there is a unique sense of community at ISC that has proved to be a crucible of a worldwide technical development. ISC is a critical part of the annual HPC cycle and has been a major contributor to the field.
Any memories of conferences past that you can share that illustrate the strength of the conference?
I have many great memories from my continuous attendance for the last 15 years, plus one involvement before that extensive duration.
I remember sitting at a table in a beer garden with Steve Wallach, Mark Seager, and others at a social event one year. An exotic belly dancer with a large snake in her performance was the entertainment for the evening.
Now, as many may know, I’m scared of snakes — like very scared. As the dancer made her way over to us, someone at our table said loudly that she should have brought a camera.
Without thinking and with a voice of anguish I said: “I should have brought a mongoose!” Steve literally fell out of his chair.
I have been honored to give the last (of three) keynote addresses every year for the last thirteen years. This came about when Hans attended a very boring panel at SC at which I was a panelist. Towards the very end, someone in the audience asked why we had not talked about multithreading.
I won’t bore you with the details, but I made the statement: “We have Intel to thank for equating the word 'hyper' with the number ‘2’ (This is related to the marketing term 'hyperthreading' that was being pushed at the time.)
The whole room burst out in laughter, some cheering, mostly due to how dull the previous hour and a half had been — the panel ended there.
Hans grabbed me a couple of minutes later and said I had to come to Heidelberg and do the same thing at ISC I had just done at SC. The rest is history.
By far the most important memory — and it is ongoing — is the large number of colleagues and friends I have formed through ISC over the many years.
I have long ago considered ISC my home on the continent, and a medium through which I have been able to mke an impact in our field.
My only regret is that I miss Hans Meuer who became over the years a colleague, a mentor, and a friend and who passed away a few years ago to the sadness of many.