On 1/30/07, Miller, Raul D <[EMAIL PROTECTED]> wrote:
> > Perhaps our different time ratios can be accounted for by
> > examining memory bus speed?
> >
> > RAM: 997 MHz, 2.00 GB, Physcal Address Extension
Dan Bron wrote:
> I have two 1 GB/266 Mhz DIMMs, and two 0.25 GB/200 MHz DIMMs
> (these were left over from the old computer; why waste 'em?).
>
> I don't know enough about hardware to infer anything from this.
It looks to me like your memory bus is about a quarter (sometimes
less) the speed of mine. Also, you've got a 2.5GHz cpu where
I'm using a 2.0GHz cpu. Looking at CPU/Memory speed ratios, the
machine you're using has something like a 5:1 bias in favor
of the CPU when compared to my machine. (CPU clock frequency
isn't really the same kind of units as memory bandwidth, but
assuming that bus width is the same in both cases, the units
cancel out when contrasting the ratios for the two machines.)
I recently acquires a pentium D dual-core which I am running under
linux. When I throw a nasty J expression at it, one which is expected
to take many, many cycles, all the work is done on a single CPU in a
single thread. The only advantage of the dual core and J in my
situation is that my computer is still responsive for other tasks
while that expression evaluates. The second CPU does little to
accelerate the process.
--
- michael dykman
- [EMAIL PROTECTED]
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
