I’ll admit it, I’m confused. I’m a big fan of The Register, but this one has me scratching my head. From a story over there posted on the 14th of May
Dell has claimed it is simplifying supercomputing by replacing complex monolithic, proprietary architecture Cray-like machines with modular ones, using racks of industry-standard components.
In fact, in one way it’s helping to complicate supercomputing, because writing parallelised code is so hard. But the massed ranks of x86 processors are blowing the Crays away and cloud-based HPC supply is on the horizon.
Um….what? Not even Cray ships a “proprietary architecture” in its primary product lines anymore, at least not in the sense of that statement (yeah, yeah, SeaStar; OK, ya got me there).
This is SGI’s model. And IBM’s (BladeCenter, not BlueGene). And HP’s. And Cray’s and…just about everyone else shipping HPC gear, with the exceptions of SiCortex and Convey, I think. Nearly every company in the market supports this revolutionary “racks of industry-standard components” model. So if Dell is going to be “blowing the [HPC industry veterans]” away with this “new” model I suggest you short Dell stock as fast as humanly possible and submit to watching 12 straight episodes of “Whale Wars” as punishment for buying it in the first place.
Also, Dell thinks GPUs might be big. Shhh.
Dell people see hybrid clusters developing with nodes equipped with multiple GPU cores as well as SMP cores. The programming task is characterised by the need to use many, many cores in parallel.
I’m not sure what happened here. I know several of the guys in the HPC business unit at Dell (US anyway), and they are reasonable folks who know their stuff and wouldn’t have spun this story this way. I also have read this particular reporter’s stuff before, and he’s usually quite good. Maybe this one was written 7 years ago and somehow just got published last week.