We checked in with Hewlett-Packard’s HPC folks recently and had an interesting conversation about their strategy and new products, and got a glimpse of their futures. One of the most noteworthy tidbits is the deal that they are working on with NEC for the Tsubame system, which will find a home at the Tokyo Institute of Technology.
It’s a pretty brawny box, with 1,400 compute nodes sporting a mix of 6-core and 8-core Intel Xeon processors along with 4,200 NVIDIA Fermi GPUs. Details on the deal and the gear can be found here and here.
Estimates put performance somewhere in the 2.4 petaflops, which should land the box somewhere in the top five or ten on the next Top500 list. This marks the first time HP has been this close to the top of the list since 2007, when they nailed down both the 4th and 5th slots with clustered blade systems in India and Sweden.
However, according to HP, this deal doesn’t signal a change in strategy – in other words, we won’t see HP chasing chart-topping deals just to cover themselves in glory. This isn’t to say that they won’t go after the biggest deals; it’s just that they have to make business sense – meaning, in our minds, deals that are profitable and not prone to the kinds of cost overruns that can plague the biggest installations.
Vendors almost always end up eating the additional costs, which tend to turn these deals from break-even propositions at best into deep pools of red ink. HP also emphasized that they are focused on using industry-standard parts rather than building one-off components that may never make it into their standard price book.
So while HP may be heading for the Top10 with Tsubame, and we’ll see them pursuing the high-end deals, we think that they will still end up doing the vast majority of their business in the less newsworthy – but much more profitable – mid-market HPC segment.