[QUOTE=Blackwood;233833]Could someone explain why Prime95 was in the prover code? I thought it was only used for testing mersenne numbers (LL), not for k*2^n-1 numbers (LLR).[/QUOTE]

Because Prime95 can PRP test k*2^n-1 numbers. :smile:
[QUOTE]11) Program can now do PRP tests of (k*b^n+c)/f. Add a line worktodo.txt that
looks like this:
PRP=k,b,n,c[,how_far_factored,tests_saved][,known_factors]
The optional how_far_factored (in bits) and tests_saved values are used
to determine if P-1 factoring prior to the PRP test would be beneficial.
The optional known_factors list is a quoted comma separated list of
known factors of k*b^n+c. [/QUOTE]
But to definitely prove that number prime, you should use other programs like LLR or PFGW.

Merfighters-
Is the PRP test much faster than LLR? I mean, is there an advantage to PRP-testing first, and LLRing the few candidates that pass a single PRP test?

-Curtis

[QUOTE=VBCurtis;233951]Is the PRP test much faster than LLR? I mean, is there an advantage to PRP-testing first, and LLRing the few candidates that pass a single PRP test?[/QUOTE]

I don't know, maybe some benchmarking would help...

[QUOTE=VBCurtis;233951]Merfighters-
Is the PRP test much faster than LLR? I mean, is there an advantage to PRP-testing first, and LLRing the few candidates that pass a single PRP test?

-Curtis[/QUOTE]
There is no speed advantage to PRP testing with Prime95 first, then finishing the proof with another program that can do a definite proof (such as LLR or PFGW)--nay, there is a bit of a disadvantage because you spend extra time doing the full proof after getting a PRP result. As far as the initial PRP test goes, though, it should take exactly the same time as an LLR test since both Prime95 and LLR are based on the same underlying gwnum libraries.

I believe the reason why unconnected uses Prime95 for his searches is because it allows automatic multithreading; you can set it to (say) use 4 cores, and it will run 1 test on each core automatically. Currently, the only way to do this with LLR is to run an LLRnet or PRPnet server to distribute the tests over multiple cores/machines.

2695-2740 complete (1 prime).

Not sure if this has already been done, but I am sieving the 2800k-3000k range, currently about 12.8k numbers remaining. Will run until same file sizes are available as currently posted (about 240 per 5k)

This has already been sieved for n<=3M to p=80000T by Primegrid. There is no need to do anymore sieving. They are also well on the way with sieving 3M<= n<=6M.

> 1 year with no progress?? So I'll take 2740-2800 range from newest sieve file posted by JimB.

[QUOTE=unconnected;312080]>So I'll take 2740-2800 range from newest sieve file posted by JimB.[/QUOTE]

Are there any new files for n<3M? I thought that he updated only those files with n>3M...

You are right, the file name is rsp3M_20110731.abcd, it's from July 2011.

They stopped sieving below rsp3M ranges because there's no need to sieve more, the optimal depth is more than passed. They are concentrating their effort on rsp7Mto9M.