The Original 'Ode To Joy' thread

[S143790]

One year after joining GIMPS:

LL Rank #800 149.726/36
TF Rank #29 129.629/109
Listed on Team Prime Rib's 'Factoring Threat Report' (not that anyone can catch them)
knocked PrimeCruncher down a notch.

Goals for the next 12 months:

Top 250 in LL
Top 10 in TF
Finding M42

[markr]

I've found factors for 3,000 exponents in LMH.

I've also just passed 250th in the GIMPS top producers list, but since I'm going the wrong direction there's no banana.

[cheesehead]

Originally Posted by markr

but since I'm going the wrong direction there's no banana.

We need a dancing/sliding banana peel.

[patrik]

I just reached 1000 P90-years.

Account ID      LL P90*  Exponents  Fact.P90  Exponents  P90 CPU
                CPU yrs  LL Tested  CPU yrs*  w/ Factor  hrs/day
--------------  -------  ---------  --------  ---------  -------
patrik          1000.74      851      30.410       75    4034.32

[Merlin]

I decided to try to find a top-5000 prime, so I started running proth.exe (stealing cycles from Prime95 of course) one evening last week. Around 2 AM that same night I got out of bed and checked the computer and could hardly believe my eyes, because it had found one (60,000+ digits) already!

[rogue]

Originally Posted by Merlin

I decided to try to find a top-5000 prime, so I started running proth.exe (stealing cycles from Prime95 of course) one evening last week. Around 2 AM that same night I got out of bed and checked the computer and could hardly believe my eyes, because it had found one (60,000+ digits) already!

FYI, Proth is the fastest primality prover for GFNs (see http://perso.wanadoo.fr/yves.gallot/primes/gfn.html). If you are not searching for a GFN, then you should use PFGW/PRP/LLR. Use it in conjunction with a number of available sieving programs such as NewPGen (the most popular) and you will find primes much more quickly. There is information on these and other useful programs on Chris Caldwell's Prime Pages.

[Merlin]

Well, I wasn't in it just for the substantial personal glory. I figured I would help a project to search for generalized Fermat primes systematically. Once I finish the ranges I reserved I might strike out on a more general search for any old big prime.

[garo]

Merlin,
Just out of curiosity, what range have you reserved? I have a range from 124000 to 125000 of 131072 reserved. I'm working on it partime using an Athlon and it is taking 9 hours per test.

[Merlin]

I meant I was looking for proth primes, of the form k.2^n+1. Brain fade. But I found another one this morning.

-----  -------------------------------- ------- ---- ---- --------------
 rank           description              digits  who year comment
-----  -------------------------------- ------- ---- ---- --------------
 2779  531*2^213721+1                     64340 g388 2005 
 3121  533*2^201797+1                     60750 g388 2005 
-----  -------------------------------- ------- ---- ---- --------------

[smh]

I meant I was looking for proth primes, of the form k.2^n+1.

Don't use proth for these kind of numbers.

First sieve a range with newpgen (removes a lot more composites than trailfactoring with proth) and then test them with prp or llr which are both a lot faster for these numbers. (oh, and do get a new prover code when you find a prime!!)

[garo]

Merlin,
Thanks for your email. smh just said what I was going to say. Since I'm using Proth.exe for GFN.

Anyway, back to on-topic discussion. My copy of Crandall and Pomerance arrived a couple of days ago