Certificate : For all indents and purposes

[Puzzle-Peter]

Originally Posted by EdH

Thanks - no hurry! It actually runs all the way to test 131 during phase 2. I would have bundled up the tmp files, but they might contain the error causing data. I think the last try was about 20 minutes in on a 10+ year old dual-core AMD machine. I have tried it on several different architectured machines including i7 and core2 quad and all seem to error the same, but I didn't keep accurate notes.

I can provide more details later today if you like.

I don't think more details will help. I was just wondering if it crashes immediately.

[EdH]

Originally Posted by Puzzle-Peter

I don't think more details will help. I was just wondering if it crashes immediately.

I must have been off in my last time check by a considerable amount. I just set it up to run again a while ago and it ran for over an hour before getting to Phase 2. So, it must have run for quite a time prior to the error message before. I canceled it this time, since I've run it several other times and it always aborted at test 131 of phase 2.

[Puzzle-Peter]

I just uploaded the certificate to factordb. Both Primo 4.1.1 and the old windows version Primo 3.0.9 produced a certificate and no crash.

EDIT: It's a certified prime now.

[EdH]

Originally Posted by Puzzle-Peter

I just uploaded the certificate to factordb. Both Primo 4.1.1 and the old windows version Primo 3.0.9 produced a certificate and no crash.

EDIT: It's a certified prime now.

Thanks! I guess I need to scour my old machines and archives for the earlier version, so I can try that myself next time...

[Puzzle-Peter]

Originally Posted by EdH

Thanks! I guess I need to scour my old machines and archives for the earlier version, so I can try that myself next time...

I guess Marcel Martin might be interested in this candidate. After all, it crashes the newest version of his software while the old versions can certify it without problems.

If you need old versions let me know.

[EdH]

Originally Posted by Puzzle-Peter

I guess Marcel Martin might be interested in this candidate. After all, it crashes the newest version of his software while the old versions can certify it without problems.

If you need old versions let me know.

I'm probably going to let him know about it and another that just showed up yesterday. This one made it all the way to the very last test and failed. I restarted it with UGS=0 and it ran for over an hour on that last test and then failed again. I cleared everything and will let it go through the cue again to a different machine. If it does the same, I'll probably let MM know about these.

I am having trouble finding my earlier versions, which is really strange - I never delete old stuff, yet every machine that is currently on has only the latest version.

I'll check back if I need 4.1.1. I'm not running any Windows.

Thanks for all the help.

[mjm]

Originally Posted by Puzzle-Peter

I guess Marcel Martin might be interested in this candidate. After all, it crashes the newest version of his software while the old versions can certify it without problems.

I would guess the same. :-)

For the first problem.
EdH stumbled on a particularly vicious discriminant. You did not hit it with the previous versions you used because the generated certificates are different but all of the Primo versions have this problem.
To fix it, append this to the .ini file:

[Setup+]
PPR=102

It increases the "Polynomial Precision Ratio" by 2% (note that 101 seems to be sufficient).

For the second problem.
"another that just showed up yesterday. This one made it all the way to the very last test and failed."
Which candidate? Which phase? How many tasks? I do not know.

Best,
mjm

[EdH]

Originally Posted by mjm

I would guess the same. :-)
...
For the second problem.
"another that just showed up yesterday. This one made it all the way to the very last test and failed."
Which candidate? Which phase? How many tasks? I do not know.

Best,
mjm

Hi mjm,

Good to see you here. I was letting this one sit while I searched my archives for an earlier version. Here's the Work Report file:

[PRIMO - Work Report]
Version=4.2.1 - LX64
WebSite=http://www.ellipsa.eu/
Work=Certification
ID=B3D5A01F6BEBF
Created=09-30-2017 09:09:07 AM

[Common]
Path=/home/math73/Math/primo/work/
Selected=1
Processed=1
Certified=0
Candidate 1=Aborted, 31603s

[Candidate 1]
Input=primo_1100000000985412192.in
Report=primo-B3D5A01F6BEBF-001.cr
Output=primo-B3D5A01F6BEBF-001.tmp
Status=ABORTED DUE TO INTERNAL ERROR 4.0.2.1

The test number that failed was:

Test 170, factoring polynomial, degree 18, bitsize 6238

The rest of the 424 tests of phase 2, all completed.

I will modify my .ini files with your addition as I run across any more troublesome candidates. Will I be able to continue from the .tmp file, or will I need to restart the run from scratch?

Thanks for all!

[mjm]

Originally Posted by EdH

I will modify my .ini files with your addition as I run across any more troublesome candidates. Will I be able to continue from the .tmp file, or will I need to restart the run from scratch?

Reloading the .tmp file should work. (You need to close Primo and to run it after having modified the Primo.ini file).

Regarding the second problem, would you send me the candidate and the data of the test 170 (in the .cr file)?

Thanks.

[EdH]

Originally Posted by mjm

Reloading the .tmp file should work. (You need to close Primo and to run it after having modified the Primo.ini file).

Regarding the second problem, would you send me the candidate and the data of the test 170 (in the .cr file)?

Thanks.

Here are the contents of the .in file:

;This is an input file for PRIMO
;Number expressed to the base 16
;The database-id of this number is 1100000000985412192, it has 3384 digits.


[Candidate]
N$=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

and .cr file [170]:

[170]
Type=EC
Gain=17
Index=4.37770
D=-3593467
H/G=180/4:45

I'll also email the files to you in a little bit.

Thanks for all.

[EdH]

Hey Pakaran,

I see we're really being hit with small PRPs again. You want me to move back down to 2k for my lower bound for a while?