[RC5] [rc5] Performance

tomdv at datatx.com tomdv at datatx.com
Wed Jan 31 12:17:40 EST 2001


Dear Basil,

agreed, you cannot compete with the top100 players. So, you are bound to
drop rankings as soon as someone new, with a dozen or so PC's joins in.
That's what's happening to me nowadays. Although my keyrate remains stable,
I'm stuck around place 16000 or so, because faster movers pass me by every
day.

But what can you do about it? I mean, if you were to set up different
rankings, how would you do it? Ask the participants to enter their number
of computers, their types, their clockspeeds, overclock speeds? Err,
difficult to set up to start with.

Then, a whole set of new calculations should be integrated as well. More
calculations mean more downtime for the statsserver. Also, see the
discussion about the averages...

And what if people cheat? They say they have only one PC where they have in
fact a dozen. We can't throw them out because they do help the project(s).
Some systems have a serial number so if we want to check that, we need to
change the cores to 'detect' it. But still, how do you know they 'belong'
to someone, ask all the suppliers in the world who bought it and then to
those people, if they sold it and to whom and so on? Quite impossible, I'm
afraid.

Basically, I think there is no practical solution for this. PC-farms will
always be stronger anyhow. Therefore, check the stats for the performance
per day for 100 people in your range or the overal performance for the 100
people in your range.

Have you also noticed, that when you do this, there are what I would call
'quiet' areas? I mean, the ranks remain stable. So, what you need to do is
to find such a range and crunch your way into it, waiting for the PC-farms
to pass by and have everyone in the range drop a few ranks.

Cheers,
Tom

--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list