Factoring by regular PRP client

The regular PRP client should have started to factor from 17.2M. Current candidates being handed out are at 17.4M, so that's 200K worth of candidates. Have any returned a factor yet?

For the work that was done by the dedicated factoring team, I think we'd expent to see about 40 factors in that 200K range, but I guess I depends whether the bounds being used are the same. I was expecting to see a collection of new names appearing in the sieve/factoring stats - all with just one result to their name, but haven't seen any that fit that bill yet.

Just to confirm, we do expect the factors to appear in the same way as if they'd been submitted by sieving or dedicated factoring and thus appear - here?

Looks like the first two have appeared.

http://www.henleyclan.co.uk/sobsieve/2009/ui/1702.htm
http://www.henleyclan.co.uk/sobsieve/2009/ui/8490.htm

Unfortunatly they are marked in the database has having "one test completed", since that's always meant a PRP test, it scores as a double check only.

Still, two factors is well down on what was seen with dedicated factoring. Does anyone know exactly why?

I think back when we were factoring we were overkilling, using way more resources and higher bounds for finding the factors than what made sense, at least I know I sure did.

Now I guess the client chooses the optimal factoring bounds for each factor which probably means that quite a lot fewer factors will be found.

Quote:

---

Originally Posted by MikeH

Looks like the first two have appeared.

Still, two factors is well down on what was seen with dedicated factoring. Does anyone know exactly why?

---

The answer lies in the default allowed memory assignment of the Prime95 client, which is too small for effective factoring.
Ref. these 2 posts, post 1 and post 2