Question about 'not verified' tests

[drew]

My computer's configured to only run 10,000,000 digit LL tests, but I just found the exponent I'm testing in the file HRF3.TXT, and it shows "Team_Italia" next to that exponent. What does 'not verified' mean, exactly? Does this mean that this exponent's been checked before, and that I've been given a double-checking assignment? I thought they were only given to slower computers on smaller exponents.

Drew

[garo]

If you can tell us the exponent in question, we could give you more details. it looks like the exponent has indeed been checked once.

It is quite likely that Team_Italia was not able to finish it or communicate to the server in time and is expired and was reassigned to you and you are now inadvertently doing the doublecheck.

Another possibility is that the first test had errors. The third possibility is that the exponent was reassigned due to a rare server failure and the reassignment was not caught in time.

[ET_]

Originally Posted by garo

If you can tell us the exponent in question, we could give you more details. it looks like the exponent has indeed been checked once.

It is quite likely that Team_Italia was not able to finish it or communicate to the server in time and is expired and was reassigned to you and you are now inadvertently doing the doublecheck.

Another possibility is that the first test had errors. The third possibility is that the exponent was reassigned due to a rare server failure and the reassignment was not caught in time.

I can confirm that our Team_Italia has had some errors in testing 10M exponents, so it is probable that Drew was rechecking one of these exponents with errors.

Luigi

[drew]

If you can tell us the exponent in question, we could give you more details. it looks like the exponent has indeed been checked once.

33232193

[dave_0273]

From hrf3.txt

33232193,Team_Italia,pippo,WX1,00010105

If you look at the last 8 numbers of that line (ie everything after the last comma), that is the error code. Ideally you want a 00000000 error code. I am assuming that George re-released this exponent for first time testing because of the scary error code.

Some additional information about this error code....
it had 5 "SUM(IMPUTS) != SUM (OUTPUTS)" errors
it had 1 "round off > 0.4" error
it had 1 "ILLEGAL SUMOUTS" error

The error code 00010105 appears 0 times in lucas.txt (verified good results)
The error code 00010105 appears 3 times in bad.txt (verified bad results)
The error code 00010105 appears 7 times (including your exponent) in hrf3.txt (a list of unverified exponents) however, nothing can be drawn from this because the exponents are unverified so the result could be good or bad.

[drew]

Originally Posted by dave_0273

From hrf3.txt

33232193,Team_Italia,pippo,WX1,00010105

If you look at the last 8 numbers of that line (ie everything after the last comma), that is the error code. Ideally you want a 00000000 error code. I am assuming that George re-released this exponent for first time testing because of the scary error code.

Some additional information about this error code....
it had 5 "SUM(IMPUTS) != SUM (OUTPUTS)" errors
it had 1 "round off > 0.4" error
it had 1 "ILLEGAL SUMOUTS" error

The error code 00010105 appears 0 times in lucas.txt (verified good results)
The error code 00010105 appears 3 times in bad.txt (verified bad results)
The error code 00010105 appears 7 times (including your exponent) in hrf3.txt (a list of unverified exponents) however, nothing can be drawn from this because the exponents are unverified so the result could be good or bad.

Thank you,

Now I feel more confident that I'm not running this test for nothing.

[99.94]

As a matter of interest, where is hrf3.txt to be found?

[PhilF]

Originally Posted by 99.94

As a matter of interest, where is hrf3.txt to be found?

Go to the GIMPS status page:

http://www.mersenne.org/status.htm

There are links at the bottom of that page to hrf3.zip and other interesting files.

[99.94]

Originally Posted by PhilF

Go to the GIMPS status page:

http://www.mersenne.org/status.htm

There are links at the bottom of that page to hrf3.zip and other interesting files.

Thanks.

[ATH]

I never checked until now but all my exponents I ever LL-tested on my pentium4 since I got it in september 2004 is in HRF3.txt under "not verified". Only 2 of them have a code different from 00000000.

From hrf3.txt in the order I did them in:

34170397,athath,p4,WZ1,00000000
33936587,athath,p4,WZ1,00000000
33762577,athath,p4,WZ1,00000000
34395763,athath,p4,WZ1,00000000
34480891,athath,p4,WZ1,02000200
34175681,athath,p4,WZ1,00000000
34578143,athath,p4,WZ1,07000700
34664863,athath,p4,WZ1,00000000
34460507,athath,p4,WZ1,00000000
34792873,athath,p4,WZ1,00000000
34922593,athath,p4,Wc1,00000000

newest one is in cleared.txt as well, the only one of them:
34922593 69 0x0173F07D5CBEA5__ 09-Dec-05 01:27 athath p4

So because I had error code on 5th and 7th I did all my exponents was moved to hrf3.txt? and why is newest one in both files?

[thechickenman]

Originally Posted by ATH

I never checked until now but all my exponents I ever LL-tested on my pentium4 since I got it in september 2004 is in HRF3.txt under "not verified". Only 2 of them have a code different from 00000000.

From hrf3.txt in the order I did them in:

34170397,athath,p4,WZ1,00000000
33936587,athath,p4,WZ1,00000000
33762577,athath,p4,WZ1,00000000
34395763,athath,p4,WZ1,00000000
34480891,athath,p4,WZ1,02000200
34175681,athath,p4,WZ1,00000000
34578143,athath,p4,WZ1,07000700
34664863,athath,p4,WZ1,00000000
34460507,athath,p4,WZ1,00000000
34792873,athath,p4,WZ1,00000000
34922593,athath,p4,Wc1,00000000

newest one is in cleared.txt as well, the only one of them:
34922593 69 0x0173F07D5CBEA5__ 09-Dec-05 01:27 athath p4

So because I had error code on 5th and 7th I did all my exponents was moved to hrf3.txt? and why is newest one in both files?

No. All exponents are put into hrf3 until a double check can be run. Since all these are over 10M digits, it'll be a long time until they're double-checked.

Though George might release the two with big errors to be checked again. Tat error code with 7's is pretty scary.