[QUOTE=rebirther;410479]I have 15 phases and was at 22% in phase 14, the ckpt looks like:

phase=14
completedStep=4
completedMaxK=26349
completedKCount=500
inProgressKCount=250

What can I do now?[/QUOTE]

I suggest that you run the entire range of k (pl_remain.txt) thru steps 14 and 15 then remove k for which a prime is found from pl_remain.txt.

[QUOTE=rogue;410487]I suggest that you run the entire range of k (pl_remain.txt) thru steps 14 and 15 then remove k for which a prime is found from pl_remain.txt.[/QUOTE]

Iam using the last version of srbsieve and try to run from the last checkpoint. What I got is:

Status (00:00:00): Started recovery at phase 13 with 2268 terms
PFGW Version 3.7.10.64BIT.20150809.Win_Dev [GWNUM 28.6]

No factoring at all, not even trivial division
Found 9 previous PRPs/Primes for distinct $a.
ABC File Processing for at most 1 Primes

Looks good so far but loosing some days. I will use the new base script next time. I have more control of it if something goes wrong and can save my last position.

[QUOTE=rebirther;410490]Iam using the last version of srbsieve and try to run from the last checkpoint. What I got is:

Status (00:00:00): Started recovery at phase 13 with 2268 terms
PFGW Version 3.7.10.64BIT.20150809.Win_Dev [GWNUM 28.6]

No factoring at all, not even trivial division
Found 9 previous PRPs/Primes for distinct $a.
ABC File Processing for at most 1 Primes

Looks good so far but loosing some days. I will use the new base script next time. I have more control of it if something goes wrong and can save my last position.[/QUOTE]

srbsieve is still much faster than the new bases script, but as you discovered, if it crashes midway thru it can be time consuming to recover.

[QUOTE=rogue;410513]srbsieve is still much faster than the new bases script, but as you discovered, if it crashes midway thru it can be time consuming to recover.[/QUOTE]

Maybe a resume feature would be sensible.

[QUOTE=henryzz;410601]Maybe a resume feature would be sensible.[/QUOTE]

Even to get it really simple: in case of crash when srbsieve is started over simply to check what last phase is over and start again.
In that case when phase is over srbsieve create single text file with phase number.
So you can manually resume at every phase ( if that case is needed)

[QUOTE=pepi37;410602]Even to get it really simple: in case of crash when srbsieve is started over simply to check what last phase is over and start again.
In that case when phase is over srbsieve create single text file with phase number.
So you can manually resume at every phase ( if that case is needed)[/QUOTE]

The resume code I have tries to resume in the middle of a phase, but it is buggy. I have no time right now to fix it.

[QUOTE=rogue;410456]Since you already tested this range to n=2500, I could just provide a list of primes up to n=10000. I think that would save us both some time.[/QUOTE]

I went thru all my results and double-checked. There are only 315 primes from n=2500 to n=10000, which I believe that I have already posted. Do you want me to recreate pl_remain.txt or do you want to do it?

Just a quick update on some bases that were started:

R598 is at n = ~13000. Last update is 547 k remaining (base started with 43726), though undoubtedly more have been removed.
R708 is at n = ~12000. Last update is 627 k remaining (started with 41120).

Taking both to n = 25,000.

S810 tested n=50k (20-50k)

93 primes found, 215 remain

Results emailed - Base released

Reserving R754 to n=100k (50-100k) for BOINC

Reserving R954 to n=100k (50-100k) for BOINC

R754 tested to n=100k (50-100k)

1 prime found, 15 remain

509*754^58046-1

Results emailed - Base released