Just before ISC late last week I got a note from insideHPC’s pal Audra in Portland (hi Audra!). Audra works with Liaison PR out there, and they handle NVIDIA’s GPU business (they are really great, by the way; if you have a company and you don’t have a great PR firm, you should be so lucky as to get them). Anyway, Audra’s email outlined the many ways that NVIDIA and GPUs in general would be part of the fabric of ISC.
Re-reading that email tonight I was struck by something that didn’t hit me on my first read of the note: the size and diversity of the partner list showing NVIDIA products or developing tools to support their hardware. From the email:
Companies either demonstrating NVIDIA Tesla GPU-based systems or offering software solutions for NVIDIA’s GPUs on the show floor include:
- Cray Inc
- E4 Computer Engineering S.p.A
- HLRS Stuttgart
- Megware Computer GmbH
- Mercury Computer Systems
- Numerical Algorithms Group (NAG)
- STFC Daresbury Laboratory
- The Portland Group
- Transtec AG
That’s a big list for a product that only relatively recently stepped onto the computation stage. Other accelerators have popped up from time to time, but it seems to me that the development of a really diverse partner ecosystem may be one of the things that could help NVIDIA’s approach stick even in the face of the threat from Intel. With companies integrating their hardware and companies developing tools to make their hardware easier to use, NVIDIA gets a dramatic force multiplier on its advertising and marketing efforts that lets it distribute marketing and sales expenses on their GPU line throughout the partner ecosystem, theoretically leaving it more money for innovation. Which would in turn make better products that would attract even more partners, and so on. I think this qualifies as a network effect, and I think it’s an important point about NVIDIA that didn’t crystalize for me until now.
If you are at ISC, you might want to poke your head in on this session on Thursday
Also, in a first this year, there will be a panel specifically on GPUs where 5 experts from industry and academia will share their perspectives on several aspects of GPU technologies, from architecture and software to tools and applications. The panel is entitled “Making Your Desktop a Supercomputer” and will take place on Thursday, June 25 from 11:30am-1pm in Hall 3.