rtomek at cis.com.pl
Thu Oct 9 10:22:53 EDT 2003
-----BEGIN PGP SIGNED MESSAGE-----
On Thu, 9 Oct 2003, Chris Samuel wrote:
> Not strictly a Beowulf as there's no Linux, but interesting nonetheless. :-)
> IrDA for the networking, 11 compute + 1 management, slower than "a mainstream
> Pentium II-class desktop PC" (they don't specify what spec).
Yes, it's nice of course. One can also build such cluster with
I myself would like to see if the performance changes after switching to
One thing that should be considered is cooling. On my iPAQ, when cpu load
gets too high for too long, the joy button warms itself. This means, cpu
is even more heated.
The other issue is power consumption. If I understand what SBP did, they
run the cluster on electricity from the wall, not from the battery. My own
observetion suggests, that running high load on battery consumes about 2-3
times more power than things like reading html files.
- From the performance side, I wonder how this compares to the following
which suggests StrongARM SA 1100 @200 is 3x faster than Pentium @166?
I was interested myself, so I ran the quick test on my own iPAQ 3630 (SA
1110 @206) and on AMD-k6-2 @475.
- -bash-2.05b# `which time` -p python /tmp/erasieve.py --limit 1000 --quiet
=> (1020 29): /usr/bin/time -p erasieve.py --limit 1000 --quiet
So, how can 12 PocketPCs be slower than 1 p2 (with no clock given at
all, but if I remember they were about 500MHz at best)? If I haven't
misunderstood something, they probably didn't tuned their experiment too
BTW, most PDA cpus lack fpu. So, while such claster may be nice to ad-hoc
password breaking, with nanoscale simulation it will be rather the
opposite, I think.
** A C programmer asked whether computer had Buddha's nature. **
** As the answer, master did "rm -rif" on the programmer's home **
** directory. And then the C programmer became enlightened... **
** Tomasz Rola mailto:tomasz_rola at bigfoot.com **
-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 5.0i for non-commercial use
-----END PGP SIGNATURE-----
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