100 MILLION DIGIT NUMBER
 

As we close in on the 10 million digit number I have been thinkng about the 100 million digit number and was wondering a couple things.
1 - what size fft would be used?
2 - when would prime 95 be able to handle numbers that large?

I would be willing to set aside a duel processor p4 2.8xeon to work on 2 of these numbers. with save files on. at several points i would be willing to move it to faster machines as I upgraded in the future. I think we are getting close to the speed to tackle large numbers. when I first started doing 10 million digit numbers the duel p3-550 took almost 14 months per number. Right now benchmarks figures 10years to finish 2 numbers on my server but I believe that number can be cut in 3 as faster processors can be swapped. if nothing else we can factor a few dozen a get some candidates.

Factoring to various levels is not too hard. There are a few of us doing factoring on billion digit candidates (see the thread and associated pages under the LMH sub-forum). Also, using the AdvancedFactor switch, folks have been using P95 to do work in the 300,000,000 exponent range.

Will GLucas work in that range?

I tried to use AdvancedFactor to factor 332192953 and prime95 keeps crashing

Prime95 doesn't work well for TFing Billion Digit candidates; it's fast but misses a lot of factors. ET_ wrote a program for Operation Billion Digits as a result. The link to download it can be found here:

[url]http://home.earthlink.net/~elevensmooth/Billion.html[/url]

[QUOTE=lpmurray]I tried to use AdvancedFactor to factor 332192953 and prime95 keeps crashing[/QUOTE]

Is that the first 100M digit Mersenne?

332 192 831 is the first (smallest) prime exponent for a 100M digit number.

I have started a new thread [URL=http://www.mersenneforum.org/showthread.php?p=31448]http://www.mersenneforum.org/showthread.php?p=31448[/URL] over in the LMH > 79.3M area for the 100M digit work.

Also started factoring work in the range 332192831 - 332199999 (will have all factored out or at 52 bits soon. ~20 min.).

[QUOTE=Uncwilly]Factoring to various levels is not too hard. There are a few of us doing factoring on billion digit candidates (see the thread and associated pages under the LMH sub-forum). Also, using the AdvancedFactor switch, folks have been using P95 to do work in the 300,000,000 exponent range.

Will GLucas work in that range?[/QUOTE]

If you are worried for MACs, now Factor3_1 works also on PowerPCs :smile:

Luigi

[QUOTE=ET_]If you are worried for MACs, now Factor3_1 works also on PowerPCs :smile:

Luigi[/QUOTE]

And Factor3_2, 25% faster, is on Billion Digits thread...

Luigi