[rescue] SMP on intel wasteful?

Bjorn Ramqvist brt at g.haggve.se
Tue Jun 25 02:43:23 CDT 2002


Chris Hedemark wrote:
> 
> > Them's fightin' words.  Do you have any IDEA of the kind of hardware
> > that Dave deals with on a daily basis, in his HOUSE even?
> 
> I've seen photos, and it is quite impressive.
> 
> Still, while that hardware is mighty studly, the dollar for dollar
> argument still trumps the CISC vs. RISC argument in terms of overall
> system performance.  The kind of people footing the bill for that exotic
> hardware are increasingly turning to cheapo CISC clusters and laughing
> all the way to the bank.

I've promised myself NOT to comment on this sort of stuff, but I feel
the need right now.

I've got several collegues at $work that thinks the same as you do.
There is a bit of truth in your saying, and there is not.
The fact is that that x86 PeeCee proc does quite a job, yes, but when
looking at RISC hardware, you don't actually focus on the proc all the
times. Check for instance the SGI Fuel and a typical common 23423GHz
PeeCee. If you look under the shell, you discover an IMMENSE I/O
capability, that completely whacks that PeeCee off the surface of the
earth. Looking at the proc, performance-wise, you'll see probably alot
less differences.
Ofcourse, this is ALL up to the applications. If your stupid application
only runs 32bit, you ain't gonna get much bang-for-the-buck with your
hotrod 64bit powerhouse. It'll even lag behind that steroid-PeeCee. If
you carefully tune the application at the development for that
particular platform, you'll get what you pay for.

When talking SMP performance, the Wintels just won't cut it.

-- 
Vnliga hlsningar/Best Regards
Bjrn Ramqvist, Hgglunds Vehicle AB



More information about the rescue mailing list