PDA

View Full Version : what should I be working on?



Silverfish
11-10-2004, 09:45 PM
I've fairly recently done a test, but that took ages (something like 70 days overall), as I've got a fairly old PC, an Athlon 1Ghz to be more precise. Anyway, now I'm doing supersecret tests, simply because they are fairly easy to set up, and I can now do a test or two a day.

I am considering what whether that is the best use of my PC, or whether Sieveing or factoring may be better. So that's what I want to ask, which would be the best thing for me to be doing now?

As I said earlier I've got an Anthlon 1Ghz, and will probably be running whatever you suggest every other day (I do Folding@home too). I was doing about 200K per second on the SB client (according to the readout in the client) with the current tests, but with supersecret with n at around 1 million I'm getting around 35-40K.

royanee
11-11-2004, 02:36 AM
The drop in K is due to CEMs and the fact that the calculations are flawed for CEM units. They inaccurately represent the increase in difficulty as n increases. I've been told that sieving is best for Athlons, but you should probably listen to whoever does a followup post. :)

Death
11-11-2004, 03:57 AM
well you can try to sieve a small range, say 100G, just to gain some new experience

Mystwalker
11-11-2004, 05:26 AM
I also suggest sieving, because

1) the range can be chosen arbitrarily, which means that you can get work for 1 hour or for 1 month - whatever you prefer and

2) Athlons are pretty good at sieving (unlike P4s), but get beaten by P4s when it comes to PRPing (the "normal" testing) and Factoring.

I'd take an even smaller range, though. A 1 GHz Athlon should do 1G (G is the unit mainly used here, it just stands for "giga" = 1 billion; it is a range of possible factors that are used to trial factor the numbers that could be prime) every hour. So maybe you should start with a range 10G for evaluation and to get familiar with the (unfortunately still manual) process.

Silverfish
11-11-2004, 05:47 PM
I've started sieving, doing 10G to start with at about 250Kp/s. I've already (at about 4.5G) gotten one non-duplicate factor, which I've submitted, as well as some duplicates and a factor out of range. Thanks for the advice everyone.

Death
11-12-2004, 06:32 AM
don't be confused if you don't see yourself in a stats.
they don't update now, and probably a few days more.

admin is on a vacations. :cheers: