Odp: [HARDWARE] 66k emails

Andrzej Hamka³o niktu at dcc.pl
Thu Oct 21 05:04:57 EDT 1999


Well, my theory was that most of the "dead" emails was caused by people that
tried
ditributed and soon got bored, so their block counts are rather modest.
According to this theory most of 145k participants unactive for last month
would fit in low segment of stats.
Otoh, "recent" move to seti at home could make that assumption inaccurate to
say at
least (or just more long-working participants leaves, than i think)
I would say that stat showing percent of people that leaved compared to
their block
count / lenght of participation would be rather interesting.
It would give a oportunity to find most "damaging" areas and help  couteract
it.
(knowing when, or how long it took to "bore" someone could give you basis to
think out a probable reason, isn't it ?)
Of course specificity of such data (it's delay and because people already
left)
could make them less useful.
Do anyone find that interesting, or is that just me ? :)


> Actually, your rank of 55k has no direct relation to what your rank
> would be against only active participants.
>
> As far as listing the info goes, I'd like to go one step further and
> provide complete stats info for 30 days and possibly 7 days as well, but
> I havn't been able to figure out a way to do it without confusing the
> heck out of everyone (well, maybe not everyone, but a lot of people
> anyway).
>
> dB!
>
> Andrzej Hamka³o wrote:
> >
> > you just spoiled my mood
> > I thought that my 55k position wasn't that bad among active participants
:(
> >
> > until now
> >
> > Anyway, there should be listed number of active e-mails since last 30
days
> > on general stats page. (or maybe 7 days also ?)
> > not every participant or team can submit their keys daily, and comparing
> > positon to number of long-dead emails isn't accurate and fun.
> >
> > p.s. I participate 465 days already and it seems i'd have to change to
> >     faster equipment, otherwise i'll be counted as unactive one ;)
> >     -and that's excuse to mail that post here :)


On the other subject:
from observavation of this list i see that best performance/price ratio
comes from celeron chips (well, even if you could get 300 of old 386 for
free,
will you pay electricity bills, and dedicate room for it ?).
Anyone thought of using 128kB cache on theese, and fit all cracking code in
it
so additional memory would become unnecesary ?
Project a board that woud fit e.g. up to 6 chips, with inteface to initially
load code,
and later exchange keys with PC. Additional benefit could be almost linear
overclocking (litlle steps). That way it could be alternative to fpga's.
I don't know if that's makes sense because i don't know if you could use
cache in such a
manner and if procesor requirements allow to build such a board without
proper laboratory.
Anyway, that's just an idea (litle bit long though :)


P.S. You can feel free to put there other procesor e.g. PowerPC

P.P.S. When you get tired of my ideas don't forget to let me know :)

--
To unsubscribe, send 'unsubscribe hardware' to majordomo at lists.distributed.net



More information about the Hardware mailing list