[Beowulf] Utility Supercomputing...

Douglas Eadline deadline at eadline.org
Mon Mar 4 08:57:30 PST 2013

>> http://www.hpcwire.com/hpcwire/2013-02-28/utility_supercomputing_heats_up.html
> well, it's HPC wire - I always assume their name is acknowledgement that
> their content is much like "HPC PR wire", often or mostly
> vendor-sponsored.
> call me ivory-tower, but this sort of thing:
>  	Cycle has seen at least two examples of real-world MPI applications
>  	that ran as much as 40 percent better on the Amazon EC2 cloud than
>  	on an internal kit that used QDR InfiniBand.
> really PISSES ME OFF.  it's insulting to the reader.

It kind of cracks me up how companies ruin their credibility
with such statements in this sector. On more than several
occasions I have offered guidance to mar-com people who were
new to HPC. I always started our conversations as follows;

"You know the rocket scientists you hear about. Well
this audience is composed of rocket scientists. They are smart and
will not believe marketing hyperbole and any plausible message
needs real performance numbers. These people and those that
support them make a living getting "it right," so speaking to
them in a technically correct fashion is important.

Indeed, if you try and BS them, you will find your company
will have reduced or no credibility in this sector. If you give
them good information, backed up with god data, they will make good
decisions, some of which may include your product. An honest
conversation is important in this sector (as it should be everywhere)."

I would say that about half of them listened and the other half
went ahead and with the classic marketing methods of
"integrating frictionless niches*" and "optimizing customized
convergence*" rather than try and understand the community.

I also want to go on record that there are many companies who "get HPC"
and do a great job addressing the needs of the market.


* http://www.codres.de/2009/07/marketing-bullshit-generator

Mailscanner: Clean

More information about the Beowulf mailing list