[QUOTE=Puzzle-Peter;438152]Done. Uploading the results file to sendspace right now. Gary will receive a message as to where to download the data.

Reserving k=23G to 25G[/QUOTE]

I got the results (residues) file but not a primes file. The file is too large for me to easily extract the primes. I will need a separate file.

[QUOTE=gd_barnes;438292]I got the results (residues) file but not a primes file. The file is too large for me to easily extract the primes. I will need a separate file.[/QUOTE]

Sorry, I forgot you're running windows. It's so easy using linux. I'll extract the primes into a separate file.

[QUOTE=Puzzle-Peter;438378]Sorry, I forgot you're running windows. It's so easy using linux. I'll extract the primes into a separate file.[/QUOTE]
FWIW, I believe Gary has Cygwin installed on his Windows box, so he should be able to run Linux commands like "grep".

In this case, the command to extract prime results from an LLR results file is:
[code]grep prime! results_file.txt > primes_file.txt[/code]
...which will write just the lines containing the text "prime!" to primes_file.txt. (If either of the file names contain spaces, put quotes "" around them.)

If you want just the prime numbers themselves, and not the full line e.g. "k*b^n+c is prime! Time : 0.0 sec.", then use this command instead:
[code]grep prime! results_file.txt | cut -f1 -d ' ' > primes_file.txt[/code]
...which passes the output through the "cut" command, which removes everything but the 1st field (-f1) using the space character as the field delimiter (-d ' ').

Grep and cut are very fast programs, so this technique will scale for even the largest files. (Unlike using find/replace in a text editor, they can work on the file as a "stream", rather than having to load the whole thing into memory and write it back out, which is the bottleneck with that approach.)

I would prefer to get the primes file instead of having to create it from a large file. I'd rather not add any more admin time to things. :smile:

[QUOTE=mdettweiler;438409]

If you want just the prime numbers themselves, and not the full line e.g. "k*b^n+c is prime! Time : 0.0 sec.", then use this command instead:
[code]grep prime! results_file.txt | cut -f1 -d ' ' > primes_file.txt[/code]...which passes the output through the "cut" command, which removes everything but the 1st field (-f1) using the space character as the field delimiter (-d ' ').
[/QUOTE]

That's interesting, I used awk for separating the primes from the rest of the line. I didn't know about the "cut" command.

The primes file is attached.

R3 tested to n=200k (150-200k) (2.147-4G)

454 primes found, 1721 remain in this range

Results emailed, Base released

Reserving R3 to n=500k (300-500k) (4-6G) for BOINC

I do not understand the following LLR output:
My input looks like this:
[code]
23566399112 59508 -1
23566399112 59543 -1
23566399112 59568 -1
[/code]and lresults.txt looks like this:
[code]
23566399112*3^59508-1 is not prime. RES64: DBDD7F48F65B3879. OLD64: A7B882071AD77B5A Time : 9.327 sec.
Iter: 94407/94407, ERROR: ROUND OFF (0.5) > 0.4
Continuing from last save file.
Unrecoverable error, Restarting with next larger FFT length...
U((N+1)/3) is coprime to N!
Time : 58.448 sec.
23566399112*3^59568-1 is not prime. RES64: B3A7CCFC753017E5. OLD64: 1AF766F55F9047AC Time : 9.324 sec.
[/code]So what does this tell me about n=59543? Prime or no prime?

Thanks
Peter

23566399112*3^59543-1 is prime! (28420 decimal digits, P = 3) Time : 15.033 sec. (i7-4930k)

I suggest that you send Jean a note to see why LLR didn't output the results to the screen.

k=23G to 25G, n=25k to 100k done. Uploading to sendspace right now