Re: Mersenne: Old files in my mprime dirs
Brian J. Beesley wrote: On Sunday 30 March 2003 21:04, you wrote: I've been running mprime on Linux for a bit less than 2 years now. I was looking at the directory on one box and found some ancient files: mE013037.001 mF320219.001 These are from a dual CPU box. The leading m suggests they're from factoring but but file sizes are 3.5 to 3.9 MB. There's been a database Very probably P-1 factoring... With a dual CPU box if you're doing P-1 on both CPUs at once you get such a file on the mprime -B -A1 process. It's just happened to me again. I checked my old prim*.log files and all those exponents were turned in. -- [EMAIL PROTECTED] - HMC UNIX Systems Manager _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Old files in my mprime dirs
Daran wrote: Actually the .001 extension would be expected... especially if running MPrime on a dual CPU box... with one instance using the -A1 switch... Using the -A switch will put an extension with the instance after it... No -A switch... no extension... OK. I've never used a Dual CPU Box. Actually... it doesn't matter whether it's a dual CPU box or not... it all depends on the -A switch... You could set up two instances on a single CPU box... and put a Time= line in the prime.ini file(s)... and have one without the -A switch running L-L tests at night... and one with the -A switch running factoring during the day... to reduce memory use... Note: ECM and P-1 use the -A switch for extensions... but the L-L tests and trial-factoring do not... L-L will only add an extension equal to iteration / X (where x is from interimfiles=X in prime.ini)... if interimfiles= is present... However... to complete the P-1s... assuming they were initiated by a L-L test... adding Pfactor=exponent,66 lines to the WORK0001.INI would do it... no ,1 is needed... since it would indicate the P-1 is complete... Not for Pfactor assignments it doesn't. Here's what undoc.txt says [...] I have verified that this is correct. A '1' in the last place causes it to choose DC limits. OK... but if the test were in progress... and not completed... MPrime has already chosen the limits... Restarting it with the DC limits... which might be different... would cause MPrime to restart the P-1 tests from scratch... A factor might be found... but if Chris didn't want to go thru the trouble... he might as well delete the files... since when double-checking gets there... the P-1 will be done... since PrimeNet doesn't know it was even started... One question: If Chris is going to finish them on a different - single CPU machine, would he have to remove the extension? Not if Chris used the -A1 switch to run the tests... but it would be easier to just remove the extension... and not use the -A switch entirely... if using a single CPU box... and only one instance of MPrime... Eric _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Old files in my mprime dirs
Chris Marble wrote: I've been running mprime on Linux for a bit less than 2 years now. I was looking at the directory on one box and found some ancient files: mE013037.001 mF320219.001 mF550789.001 mF614243.001 mF687599.001 These are from a dual CPU box. The leading m suggests they're from factoring but but file sizes are 3.5 to 3.9 MB. These are from P-1 runs... not factoring... The first letter will determine the test... E = ECM... M = P-1... P = Trial-Factoring OR L-L... Trial-factoring files are always 32 bytes... The fact that they are 3.5 to 3.9 MB indicates... that they all are P-1 tests in progress... and not completed... Prime95 (and MPrime) will delete P-1 files automatically if run as part of the L-L test... If they were P-1 tests run by themselves... and completed... they would be from 1.75 to 2 MB... You are correct... however... in assuming that E=14 and F=15... in this case... To check what happened to these exponents and the results... and whether they were reported to PrimeNet... you will have to look at the RESU0001.TXT and PRIM0001.LOG... RESU0001.TXT will give whatever results Prime95/Mprime came up with... and PRIM0001.LOG will tell what results... in any... were reported to PrimeNet... What happened that these files were left around a year ago? If the P-1 was part of an L-L test... there could've been a crash or something... that prevented Prime95/Mprime from completing the tests... and therefore not deleting the files... Eric _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Old files in my mprime dirs
On Sun, Mar 30, 2003 at 01:04:10PM -0800, Chris Marble wrote: I've been running mprime on Linux for a bit less than 2 years now. I was looking at the directory on one box and found some ancient files: mE013037.001 mF320219.001 mF550789.001 mF614243.001 mF687599.001 These are from a dual CPU box. The leading m suggests they're from factoring but but file sizes are 3.5 to 3.9 MB... They'll be P-1 save files. There's been a database sync so I don't have results for 14013037 (assuming E=14, F=15) on my page. Where would I check to see that all these exponents had been completed so I could delete these old files? They've probably had a first time test, but are unlikely to have been DCed. The file ftp://mersenne.org/lucas_v.zip will tell you, but I don't one which is up-to-date, and it's a multimegabyte download. I do have the the latest pminus1.txt file, though, and these exponents aren't listed. What happened that these files were left around a year ago? They shouldn't have an extension, so I'm gessing. Backup and restore? Recovered from a damaged filesystem? My recommendation would be to complete the P-1 as though they were DCs. Just remove the extensions and put Pfactor=exponent,66,1 into your P-1 factory's worktodo file. -- [EMAIL PROTECTED] - HMC UNIX Systems Manager _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Old files in my mprime dirs
At 21:14:50, Sunday, 3/30/03, Daran wrote: On Sun, Mar 30, 2003 at 01:04:10PM -0800, Chris Marble wrote: I've been running mprime on Linux for a bit less than 2 years now. I was looking at the directory on one box and found some ancient files: mE013037.001 mF320219.001 mF550789.001 mF614243.001 mF687599.001 They shouldn't have an extension, so I'm gessing. Backup and restore? Recovered from a damaged filesystem? I know that files named pXXX.001 are interim files generated during the LL test if you have turned this option on in prime.ini (see undoc.txt for more on this feature). Perhaps, these files are P-1 interim files? Just a guess. -- Nick Glover [EMAIL PROTECTED] It's good to be open-minded, but not so open that your brains fall out. - Jacob Needleman _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Old files in my mprime dirs
Daran wrote: Chris Marble wrote: What happened that these files were left around a year ago? They shouldn't have an extension, so I'm gessing. Backup and restore? Recovered from a damaged filesystem? My recommendation would be to complete the P-1 as though they were DCs. Just remove the extensions and put Pfactor=exponent,66,1 into your P-1 factory's worktodo file. Actually the .001 extension would be expected... especially if running MPrime on a dual CPU box... with one instance using the -A1 switch... Using the -A switch will put an extension with the instance after it... No -A switch... no extension... However... to complete the P-1s... assuming they were initiated by a L-L test... adding Pfactor=exponent,66 lines to the WORK0001.INI would do it... no ,1 is needed... since it would indicate the P-1 is complete... Eric _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Old files in my mprime dirs
On Sun, Mar 30, 2003 at 09:58:15PM -0800, Eric Hahn wrote: Daran wrote: Chris Marble wrote: Actually the .001 extension would be expected... especially if running MPrime on a dual CPU box... with one instance using the -A1 switch... Using the -A switch will put an extension with the instance after it... No -A switch... no extension... OK. I've never used a Dual CPU Box. However... to complete the P-1s... assuming they were initiated by a L-L test... adding Pfactor=exponent,66 lines to the WORK0001.INI would do it... no ,1 is needed... since it would indicate the P-1 is complete... Not for Pfactor assignments it doesn't. Here's what undoc.txt says #You can do P-1 factoring by adding lines to worktodo.ini: # Pfactor=exponent,how_far_factored,has_been_LL_tested_once #For example, Pfactor=1157,64,0 I have verified that this is correct. A '1' in the last place causes it to choose DC limits. One question: If Chris is going to finish them on a different - single CPU machine, would he have to remove the extension? Eric Daran _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
