[Beowulf] Given all the recent infrastructure talk...

[Mark Hahn]

> ...it looks like some folks still subscribe to the "racks of desktops"
> model.  See

that's great news!

the worst thing about academic HPC in recent years 
is the incredibly risk aversion and bureaucratic sclerosis. 
when an organization only ever considers buying name-brand products, 
it's toast, game over, no coming back.

it's a testament to organizations like FB and Google
who know themselves in enough of a competition to stay sharp:
they have always built their own hardware.  they've merely graduated
from buying parts from Frys to custom-manufactured designed unhampered
by the grotesque priorities of "Enterprise" server vendors.
cutting out the middle man doesn't hurt either, I'm sure.

I think in some sense, the problem is that in academic HPC organizations,
decisions are typically made by academics recruited to be management,
and they have either a high fear/expectation of failure or a low expectation
in being able to fix problems that do arise (or both).  it's crippling,
and being emotional, prevents such organizations from considering how to 
rationally estimate the risks, and to design the process to manage it.

in a sense, beowulf has been corrupted by its own success.
hacking (in the classic sense) is inherently risky.

bitter?  sure.  to me Canadian HPC is on the verge of extinction,
partly because of this issue.

sorry for the vent!