C++ programming (was Newbie Alert: Beginning parallel program ming with Scyld)
shaeffer at neuralscape.com
Fri Oct 18 22:07:16 EDT 2002
On Fri, Oct 18, 2002 at 03:31:58PM -0700, Jim Lux wrote:
> I think a lot of the discussion really centers around the question:
> Is it cheaper to
> 1) use more faster processors and accept some inefficiency in the code
> 2) use more programmer labor and make more "efficient" use of the processor(s)
> I maintain, philosophically, that since processor power is continuously
> getting cheaper than programmer power, the trend should be to less "tuning"
> and more use of massive processing (which is what Beowulfs are all about in
> the first place, or we'd be agitating for personal Crays, right?).
I apologize for consuming the bait, but...
Of course, this argument is embraced by business managers and taught at
universities. But, in practice, it often turns out to be an
oversimplification and leads to very costly problems.
For example, I know of a local company that produced a Java based Enterprise
application that was so hyped they got like 50 million in venture funding.
Well, to make a long story short, once the product was put into real
corporate networks--it failed miserably. And no easy fixes were to be found,
leading to the company going broke and laying off all their wiz bang Java
The point is that with every layer of abstraction that you don't have
control of, your overall system risk goes up. This is the other side of the
equation. And, as has often been discovered after the fact, fixing system
level failures that are sitting on top of layer after layer of abstractions
the development team doesn't control, is often times impossible.
Each project needs to be evaluated based on it's particulars. In many
instances, folks are choosing the c language, because it affords the
development team the advantage of projecting most risk factors into their
sphere of control. You trade-off labor costs up front for a leveraged
mitigation of risk factors later on.
Neuralscape; Santa Cruz, Ca. 95060
shaeffer at neuralscape.com http://www.neuralscape.com
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
More information about the Beowulf