Whether you need technical insight, networking time or career development, a guide to key events of the high-performance computing calendar is essential, says Andrew Jones.
As I've noted in previous columns, high-performance computing (HPC) is an area that is growing rapidly as it becomes increasingly important to businesses — and yet many aspects of HPC still need a specialist. Making use of the HPC community can help you keep up with the fast changes in technology and novel methods of performance programming.
Whatever your scale of technical computing, from multicore workstations to multi-thousand-node supercomputers, getting involved with the active HPC community can help you with your parallel computing goals. Online resources can help, but by far the most effective way of benefiting from the wider HPC community is by participating at the right events.
In fact, people starting out in HPC, whether in a commercial capacity or as new researcher, often ask me which conferences they should attend. So here are some suggestions based on the events I, or one of my colleagues, attend regularly.
First, there are the two all-in-one conferences, of which the biggest is the annual Supercomputing Conference, colloquially known as SC. It is held each November in the US and offers a large technical programme and a huge exhibition — think major motor show or airshow size.
It is so all-encompassing that some people have commented that the absence of any vendor or major HPC centre is in itself news in the community.
As well as the official conference, there is a whole side conference taking place in the surrounding hotels, as vendors run non-stop briefings for key customers, and the heads of supercomputer centres assess the market. There is no single bigger week in the HPC calendar.
The International Supercomputing Conference (ISC), held each June in Germany looks similar to SC, just a little smaller, with about 2,000 attendees as opposed to about 10,000. But some people think ISC is a friendlier event because of its smaller size and different type of location. I find the technical programme at ISC often holds more interest than that at SC. Certainly the potential for ad hoc networking is very good, as opposed to the aggressively planned networking at SC.
Networking and case studies
Then there are the smaller workshops, typically with 100 to 200 attendees, which place a greater focus on networking opportunities and presenting case studies, along with updates from major HPC users, national labs and vendors.
Of these types of event, the two that stand out for me are the HPC User Forum, hosted twice a year in the US, and the HPCC conference held in Newport, Rhode Island each March, also known as the Newport event. If using or managing HPC is your main role, then these two workshops are great for finding out what similar organisations are doing, and to pick up tips and ideas.
For those new to HPC, perhaps considering buying your first large cluster or formalising HPC as an activity in your organisation, then these two events are approachable and much less daunting than the large SC and ISC conferences.
Vendor pitch or technical
Nearly all conferences in HPC, as with any other field, will have their share of vendor pitches. But some of the better conferences are closely allied to specific vendors.
Two that I have regularly attended are: the annual Cray User Group conference, which is independent from Cray but focuses on its products and users, although much of the technical material is applicable to any large supercomputer; and the Intel HPC Roundtable. This event is organised by Intel in the US and Europe each year, inviting key HPC users.
There are also similar events focused on SGI, IBM and HP.
At the other end of the spectrum are those conferences that are predominantly about the technical programme, such as the SIAM Parallel Processing or CSE conferences. These are great for career development if you are involved in research into HPC technology, a programmer of HPC systems, or a scientist making regular use of HPC.
Finally, there are the workshops that are niche nuggets. These are smaller workshops that centre on a specific theme. If that theme or the resulting audience is relevant to you, then these can be very productive events. These are often organised by national labs, major HPC centres, or groupings of end-users.
They may be focused on an industry — for example, aerospace or oil and gas — or technology, such as many-core computing or exascale software challenges. These workshops bring together users in a given field to discuss how to do HPC better, or to learn about the challenges and opportunities coming from future HPC technology.
Use the community
So that is my run-through of the various types of events for exploiting the wider HPC community for your benefit. I may have missed out your favourite conference, in which case I would like to hear about it.
It may be that your organisation finds it hard to justify attending conferences, but the money invested in community participation might help you get better efficiency from the far larger budget allocated to the rest of your supercomputing.
Many people have rightly remarked that the HPC community really is that — a community — and that there is still a relatively high degree of connection between the various practitioners. In other words, despite its growing size and global reach, it feels like a small community. People know each other.
Consequently, networking, whether technical or commercial, goes a long way to helping your business.
As vice president of HPC at the Numerical Algorithms Group, Andrew Jones leads the company's HPC services and consulting business, providing expertise in parallel, scalable and robust software development. Jones is well known in the supercomputing community. He is a former head of HPC at the University of Manchester and has more than 10 years' experience in HPC as an end user.