I have attended every Supercomputing conference since 1992 in Minneapolis, and boy things have changed. First of all, the conference has gotten so big that for many years now, it would not have even fit in the Minneapolis Convention Center. It has been about a month since the show, and I have some thoughts about what I saw and heard.
I think much of how I feel can be summed up in a single phase: Nothing disruptive. Yes, vendors showed new technologies from storage appliances, interconnects, faster CPUs, new SSDs and so on. I do not think any of these technologies qualify as something that will change the face of high performance computing. Infiniband was one such technology when it came out. It allowed the industry to go from large SMP machines to commodity clusters, and it was truly innovative in my opinion. I am not sure if the industry has done a whole lot to change the face of HPC since then.
This is not to say there have not been technological progress almost every area (e.g., CPUs, memory bandwidth, file systems and SSDs), but progress and massive changes are two different things. Major changes do not happen overnight, of course, nor do they happen every year, and they almost never happen when there has been limited investment in technologies because of a recession. Additionally, there might be a cycle to innovation in this field that happens every 10 to 15 years, looking back from the 1960s to today. But that is a different story for another day. My hope is that some time in the next two to three years something comes out that changes things again in a very positive way. I am not sure what that will be, as there are many possibilities, but I hope it is soon.
Labels: supercomputing,HPC,SSD,storage technologies
posted by: Henry Newman