Distributed computing, Was: "Device not ready" (CDBurn)

As Ben Shimmin <bas@xxxxxxxxxxxxxxxxx> wrote:

> druck <news@xxxxxxxxxxxxxxxx>:
> > On 13 Jan 2006 Ben Shimmin <bas@xxxxxxxxxxxxxxxxx> wrote:
> >> [Jan 13 16:15:57 UTC] RC5-72: Summary: 2 packets (2.00 stats units)
> >> 0.00:03:42.92 - [36.43 Mkeys/s]
> >>
> >> (on a dual 2.5GHz Power Mac G5 -- the 2004 model, not the new one).
> >>
> >> Perhaps you could paste the results from your Iyonix? :)
> >
> > I'm surprised anyone is running that, its pretty pointless. I stupidly ran
> > it on my Risc PC for a couple of years slowing the machine down
> > uncessessarily.

A remark that nicely fits in with the CMT/PMT debate: On RISC OS it is
impossible to run the distributed.net client without slowing down other
work significantly. On my Linux machine on the other hand, there is no
noticeable impact on usability, despite the fact that the client runs at
top speed whenever the machine is idle.

Preemptive Multitasking, complete with priority levels for processes,
clearly has the advantage in this case.

> > If you've got spare CPU cycles on a non-RISC OS machine for which clients
> > are available, why not do something useful with them such as cancer
> > research, or at the very least find yourself some aliens.
>
> I absolutely, wholeheartedly agree (a first for us, I should think!). I
> only downloaded the RC-72 client to see how fast my machine would do it --
> I've never used it otherwise, and I've never taken part in any of the other
> distributed.net challenges, which, beyond being a (fairly obvious) proof of
> technology, serve absolutely no purpose.

Yes indeed. I'm getting rather disheartened with it myself. Pros:

- The client can serve as a nice benchmark to compare processor
performance for small, integer-only programs. Nowadays mainly showing
that the ARM, in any of it's guises, lags sadly behind any other
processor used in a desktop machine.

- Demonstrating that short encryption keys are no longer safe enough.
In particular, it was shown that the DES encryption does not provide
any practical safety any more.

- Bringing Acorn/RISC OS to the attention of the computing public. But
only when we find the code. Currently the Acorn Users Team is
approaching the Top-100 of teams, but does this mostly by cracking keys
on non-RISC OS machines. See http://www.acornusers.org/rc5/ and
http://stats.distributed.net/team/tlist.php?project_id=8&low=100&limit=100 .

- As mentioned: Proof of technology. Projects like Folding@Home would
not have existed if distributed.net had not explored the field.

Cons:

- No direct practical purpose.

- Only a slim chance of winning the prize.

> The Folding@Home project, on the other hand, is something which I believe to
> be extremely worthwhile, and in which I do partake:
>
> <URL:http://folding.stanford.edu/>

Looks interesting. Only runs on Windows, Linux (on i86) or Mac OS X
though. Are there any other useful distributed computing projects out
there?

> I'll leave finding aliens for Mr Vigay.

My own main objection against seti@home is this: Humankind has been
sending radio signals for about 100 years. Really powerful signals
(more than 1 Mega Watt) for about 50 years. We hope to be sending
signals for a long time, at the very least for 2000 years.

This means that it is highly likely that any aliens sending signals we
can detect, will have been sending them a lot longer than we have. This
in turn means that, assuming the same rate of development of any
aliens, it is highly likely they will be much more advanced then us.
This is made worse by the time the signals will have travelled (probably
more than several thousand years).

So these potential aliens will be a lot more advanced than us. In that
case, I don't want to find them. More importantly, I don't want them to
find us.

Projecting from human history, Very Bad Things happen to societies that
are visited by foreigners that are more technologically advanced.
Usually involving genocide and enslavement.

So I'm probably going to stick with contributing to medicine :-).

--
Erik Groenhuis http://www.xs4all.nl/~erikgrnh
Home of RCS for RISC OS v5.7.1.2
StrongARM / RISC OS 4.02 || AMD K7-7 (Duron) / Red Hat Linux 9.2
.

Relevant Pages

  • Re: Manufacture vs. Artifact?
    ... thoroughly investigated the cube and tested hypotheses of manufacture. ... You already know that a highly symmetrical polished granite cube can ... Certain types of signals, if discovered actually coming from outer ... was contrived to be something that we would expect aliens would create ...
    (talk.origins)
  • Re: chances are greater of this being alien contact than not; signal in Pisces and Aries
    ... Reports of SETI@home Extraterrestrial Signal Highly Exaggerated ... billion potential signals, this candidate, labeled SHGb02+14a, was ... > Exciting also in that the way we do science in the future will be vastly ... > are true advanced aliens should be these: ...
    (sci.physics)
  • Re: chances are greater of this being alien contact than not; signal in Pisces and Aries
    ... Reports of SETI@home Extraterrestrial Signal Highly Exaggerated ... billion potential signals, this candidate, labeled SHGb02+14a, was ... > Exciting also in that the way we do science in the future will be vastly ... > are true advanced aliens should be these: ...
    (sci.astro)
  • seti...
    ... If aliens really are conversing, ... listening for such signals, says Claudio Maccone, co-chair of the SETI ... A few people have been "preaching the KLT" since the early 1980s but ...
    (soc.culture.romanian)
  • Re: chances are greater of this being alien contact than not; signal in Pisces and Aries
    ... Reports of SETI@home Extraterrestrial Signal Highly Exaggerated ... billion potential signals, this candidate, labeled SHGb02+14a, was ... > Exciting also in that the way we do science in the future will be vastly ... > are true advanced aliens should be these: ...
    (sci.logic)
Loading