> 
> I guess it is more about being part of the corporate culture. Always 
> solving the big problems.
> 
> You look at their website and they aren't solving little problems. They 
> are doing big lustre clusters..
> 
> They are good at all things scale.. more CPU to solve problems, more io, 
> more network, more disk, more memory..
> 
> Scale though meaning through hardware and software architectures.
>

Yes, this makes sense. You will hear in the industry that there are very good
benchmarks, etc. _Everybody_ I talk to say that they can do "this" and "that"
and scale very well, etc. I found a post by [company] on LinkedIn where they
showed computational results scaling very well to over 20,000 cores. They were
bragging that their method was scaling well while taking about 4 minutes per
timestep to advance the oslution. It was almost a joke to read... Knowing what
was expected of the method in terms of computational demands, their calculation
was not achieving any real scale in my book. (End rant.)

 
> I have met 100's of cray people over the years, they never answer a 
> question.. the answer is always maybe..
>

Ha! I can relate to that. The answer is a combination of "maybe" and "it
depends." And I value much more the opinions of those people to people with
arbitrary claims. If CRAY's attitude is this, it is no surprise they are in
business today.

We went through a "cottage industry" stage in supercomputing, where even Apple
had come to me with a RISC box and asked me to run benchmarks with my solvers
to see if we would be benefiting from buying hardware (a cluster) from them.
They were using compilers that had come form IBM, and I think they were
trying to compete with entities like DELL, who tried to enter the supercomputer
market as a vendor with dirt-cheap hardware. That was a pain in the neck to
CRAY, I am sure, and likely contributed to their near disappearence. But look
who does supercomputing better now...

Mellanox did it right by hammering on the interconnect side with Infiniband.
That is the part of the cottage industry that actually gave back value.


Thanks for your thoughts. If you have more specific examples of CRAY's work
based on your interactions, I'd love to hear about. (Send links)