Re: Although Sun had adapters for CompactPCI as far back as 1997
There are several topics "in play" here.
CompatPCI busses, which you seem to think makes some huge difference when it comes to re-configurable clusters of multiprocessor servers, and scalable multiprocessor systems.
Do you think SUN has any market at all? Or are you convinced that they're a startup hoping to someday sell their first product?
I think you have some inkling that they actually do already sell products, and that their market presence includes being one of, if not the major player, in large, scalable systems.
We all know how behind single Sparc CPUs have gotten, so how, exactly, do you think SUN has managed to keep being "the one to beat" in large systems?
The answer is that, partly through necessity, they have become the leaders in reconfigurable multiprocessor systems configured into clusters.
Now, from the land of "netburst", "peformance enhancing 820 chipsets", and "the first 12" wafer FAB", comes "blades". Where, typically, Intel and some partners are suggesting that their particular version of a crummy or inappropriate use of a technology is something fabulous because they're reduced to shipping it.
Using a fairly cheap, old, bus, that's been given some credibility as a high performance standard through SUN's use of it for high end I/O as a processor bus is a pretty lame thing to do, especially when the competition is already shipping far superior "plug in processor" solutions, but it's classic "big lie" stuff from Intel.
Just because you have a new name and the wrong bus for something doesn't mean you invented it, and doesn't make it something new and better. |
