distribution of remaining candidates
 

paul, could you give us a table with

range of 1000 | remaining candidates
700 | 75
701 | 69
....
999 | 81

you knpw what i mean?
the more remaining candidates in a range, the bigger the chance of a prime!

thommy

Here it is:

moved to [URL=http://www.mersenneforum.org/showthread.php?s=&threadid=543]here[/URL]

thanks
 

the tops are:

707 89
748 83
849 81
934 80


the bottoms:

852 45
860 45
749 46
901 47


i bet there is a prime in at least one block of the top 4. no doubt about it.

I have some doubt! I just hope we find another prime before n=1,000,000. Even two!

Everyone is free to "cherry pick" ranges. The time taken to test the smaller numbers is less.

:help: Perhaps some people will join in and bring with them their beginners luck!

I will be releasing all remaining candidate prime numbers in blocks of a 1000 range.

At n=700,000 a block of a 1000 range will take about 4-5 days on a P4 @ 2.4GHz -- take into consideration the above table. I will try out a high block after I have finshed my current block.

I will maintain the above table so that it reflects what is left :geek:

Other methods to sieve
 

Once some of the lower candidates have been removed sieving again can reduce few more candidates. Also other methods should be tried to factor these numbers like P-1 and ECM.

Good Luck with the project!
I also hope you find 2 primes!

Citrix
:cool: :cool: :cool:

[QUOTE]Once some of the lower candidates have been removed sieving again can reduce few more candidates.[/QUOTE]

We are through with sieving. There is no point sieving further because we reached the limit of effectiveness of sieving.

[QUOTE]Also other methods should be tried to factor these numbers like P-1 and ECM. [/QUOTE]

If the were a program I could run against the remaining candates I would use it. Is there a P-1 program that can cope with 200,000+ digit numbers? If so where is it and how do I use it?

P-1
 

Paul,
You will have to write one for yourself. The algorithm of p-1 is really simple but to optimize it for large numbers is the tricky part. The GIPMS and the SOB project use this code, you could ask them to share their source with you or modify their p-1 client for you.

Good luck!

Citrix

:cool: :cool: :cool:

[QUOTE][i]Originally posted by paulunderwood [/i]
[B]Is there a P-1 program that can cope with 200,000+ digit numbers? If so where is it and how do I use it? [/B][/QUOTE]

The GMP-ECM program will do P-1 and P+1 for numbers of this size. For numbers with millions of digits you will need to recompile following the instructions. But I just checked a number with 220000 digits using the GMP-ECM5 beta that is in the files section for the Yahoo PrimeNumbers Group. This is the version used by ElevenSmooth, so you could also get it from the ElevenSmooth [URL=http://home.earthlink.net/~elevensmooth/ElevenDown.html]Download Page[/URL].

optimized?
 

wblipp,
Is the code optamized for such large numbers? Specially for numbers base 2.

Citrix
:cool: :cool: :cool:

How do I perform p-1 only with ecm?

wblipp, how did you run your run? I want to input a file of numbers.