[QUOTE=rebirther;448720]But its sieving not testing...[/QUOTE]

hi,
how do you mean that?
i do know the difference between sieveing and prp-testing
i reserved S742 for to sieve for BOINC as i wrote before
i think that after the sieve is done by me you will reserve it for
prp testing with BOINC
AFAIK you did the sieve (not deep enough) by yourself and loaded the remaining candidates
to the BOINC server
so my sieveing work is meaningless

hi,
to gd_barnes:
please do cancel my reservation for S770

[QUOTE=lalera;448723]hi,
how do you mean that?
i do know the difference between sieveing and prp-testing
i reserved S742 for to sieve for BOINC as i wrote before
i think that after the sieve is done by me you will reserve it for
prp testing with BOINC
AFAIK you did the sieve (not deep enough) by yourself and loaded the remaining candidates
to the BOINC server
so my sieveing work is meaningless[/QUOTE]

I have taken the sievefile from the reservation table. I cant see any problem with the amount of work for 25-100k.

[QUOTE=rebirther;448730]I have taken the sievefile from the reservation table. I cant see any problem with the amount of work for 25-100k.[/QUOTE]

hi,
it is only sieved to p=100e9

[QUOTE=lalera;448727]hi,
to gd_barnes:
please do cancel my reservation for S770[/QUOTE]

Lalera, I'm sorry if you were upset by the exchange with Reb. We have greatly appreciated the amount of sieving that you have done for the project. I think there was some misunderstanding at the sequence of events and the nature of BOINC. I hope that you will consider continuing to work on CRUS. You can work outside of BOINC on your own efforts if you would like.

Here are my suggestions to avoid this problem in the future:

Reb, 1-2 years ago we had a searcher do some sieving for n=25K-100K on many bases to only P=100G as a starting point. The thinking then was that I would post them and whomever reserved them would sieve them further before beginning testing. Lalera has been doing additional sieving on some of these files to bring their sieve depth up to where they need to be. This has saved BOINC from testing many composite numbers. I know that BOINC does not care very much if it does many additional tests but technically its resourses can be better utilized by searching files that are sieved to a proper depth. I believe it was you who started the sieving for BOINC thread so here is my suggestion for the future. If you see a sieve file on the reservations page and you are wanting to test it, see if there is an in-process reservation in that thread to sieve the file further. If there is consider reserving something else until that sieving effort is complete. The bottom line is: I do not suggest testing any file that is only sieved to P=100G.

Lalera, if you do decide to sieve in the future, see if there is already a file for the base that you are considering on the reservations page. If so consider sieving something else. This will avoid this kind of problem. I realize that many files are under-sieved but if someone reserves a base for testing it is up to him to make sure that the file is sieved deep enough. If BOINC wishes to test under-sieved files that is their perrogative although I do not suggest it.

Edit note: I have now posted the more deeply sieved files for both R742 and S742 on the reservations pages.


Gary

[QUOTE=gd_barnes;448733]Lalera, I'm sorry if you were upset by the exchange with Reb. We have greatly appreciated the amount of sieving that you have done for the project. I think there was some misunderstanding at the sequence of events and the nature of BOINC. I hope that you will consider continuing to work on CRUS. You can work outside of BOINC on your own efforts if you would like.

Here are my suggestions to avoid this problem in the future:

Reb, 1-2 years ago we had a searcher do some sieving for n=25K-100K on many bases to only P=100G as a starting point. The thinking then was that I would post them and whomever reserved them would sieve them further before beginning testing. Lalera has been doing additional sieving on some of these files to bring their sieve depth up to where they need to be. This has saved BOINC from testing many composite numbers. I know that BOINC does not care very much if it does many additional tests but technically its resourses can be better utilized by searching files that are sieved to a proper depth. I believe it was you who started the sieving for BOINC thread so here is my suggestion for the future. If you see a sieve file on the reservations page and you are wanting to test it, see if there is an in-process reservation in that thread to sieve the file further. If there is consider reserving something else until that sieving effort is complete. The bottom line is: I do not suggest testing any file that is only sieved to P=100G.

Lalera, if you do decide to sieve in the future, see if there is already a file for the base that you are considering on the reservations page. If so consider sieving something else. This will avoid this kind of problem. I realize that many files are under-sieved but if someone reserves a base for testing it is up to him to make sure that the file is sieved deep enough. If BOINC wishes to test under-sieved files that is their perrogative although I do not suggest it.

Edit note: I have now posted the more deeply sieved files for both R742 and S742 on the reservations pages.


Gary[/QUOTE]

ok, I understand, whats the difference now? Only less tests? I will run the test with the old sievefile to 50k, then I can take the rest from the new sievefile if its possible.

[QUOTE=rebirther;448735]ok, I understand, whats the difference now? Only less tests? I will run the test with the old sievefile to 50k, then I can take the rest from the new sievefile if its possible.[/QUOTE]

Yes there will be many fewer tests and...that is a good idea to go ahead and test the old file to n=50K, delete tests from your server for n>50K, and then load the new file for n=50K-100K. I don't know how your server handles tests for k's that already have a prime. When you load the new file, you may need to manually remove all tests for k's where there was a prime for n=25K-50K.

Lalera has done a lot of good sieving both for BOINC and potentially for others outside of BOINC to reserve work for themselves. I personally have used at least one of his files when testing. It is my hope that the misunderstanding can be resolved to his satisfaction.

Reserving R982 to n=100k (25-100k) for BOINC

R798 tested to n=200k (100-200k)

2 primes found, 4 remain

302*798^104367-1
322*798^104936-1

Results emailed - Base released

While doing unrelated work i've found 39939*1030^25030-1 is 3-PRP!

S856 tested to n=100k (25-100k)

65 primes found, 103 remain

Results emailed - Base released