Mersenne.ca discrepancies

[ixfd64]

I recently noticed a few discrepancies with the Mersenne.ca database:

1. The entry for M87940961 shows that the minimum P-1 bounds are not calculated. I haven't seen this for any other exponent. Is there any reason for this?

2. The entry for M25153 incorrectly lists 95798344469709111399327318847787476387212693634994913715944997998172541010266957318247461159 as a prime factor. It's actually the product of the four currently known factors.

[heliosh]

Originally Posted by ixfd64

2. The entry for M25153 incorrectly lists 95798344469709111399327318847787476387212693634994913715944997998172541010266957318247461159 as a prime factor. It's actually the product of the four currently known factors.

It probably was accepted as a new factor because it was used in a PRP test:
https://www.mersenne.org/report_expo...exp_hi=&full=1

[Batalov]

Same thing appears to be happening in

http://www.mersenne.ca/exponent/7508981

196390556749512365885553091471517315772448332119011812799224967043247355807885758998536981636538241093908075798179275216900756503990104076135107768463 
is not a prime factor

http://www.mersenne.ca/exponent/111269

2483611811780967729450526041810356652852620293014220023255782117303084384090070203366756119
97741873979710938487033764107284665796924065785506374732657395204556552490513667177770940383463688455655450791
92931654532319876438832186923672422729542099694851688102212583901293792387759635634014037818755101113436948191796372627855844302598713
are not prime factors

[James Heinrich]

Originally Posted by ixfd64

I recently noticed a few discrepancies with the Mersenne.ca database

Please feel free to send me an email or PM if you notice any issues as I don't frequently browse the forum and wouldn't have seen this thread (thanks to GP2 for bringing it to my attention).

Originally Posted by ixfd64

entry for M25153 incorrectly lists 957...159 as a prime factor

When new factors are submitted they are immediately checked for primality, but if they exceed a certain size (I believe I set it to ~200 bits) they are temporarily accepted and set aside for me to check manually. Unfortunately I was on vacation for a couple weeks and a typo in my verification code prevented the job from running so a few composites slipped in. I'm running a double-check now and removing composite factors. So far I have just found 5 factors for the three exponents mentioned in this thread (M25153, M111269 (x3), M7508981). If I find more I'll purge them too.

Originally Posted by ixfd64

The entry for M87940961 shows that the minimum P-1 bounds are not calculated. I haven't seen this for any other exponent. Is there any reason for this?

A side-effect of the above-mentioned typo in the cron job that is tasked with such things. Once it finished the backlog the bounds should show up correctly.

[Prime95]

Even a one or two thousand bit PRP test should be pretty near instantaneous. Alternatively, attempting to divide by known factors would be very quick. Once setup, it would save you from manually handling big composites.

[James Heinrich]

Both fair points. I'll see if I can work those ideas into the codeflow.

[Uncwilly]

Originally Posted by Uncwilly

I am throwing a bunch of P-1 at 79777211 (I went to mersenne.ca and asked for the worktodo entry with a 5.5% chance of finding a factor.)

When doing the cut and paste of the exponent number (from the bit highlighted in the attachment) into the P-1 form, it lopped off the final '1'. I found this out because 7977721 already had a factor reported (which I rediscovered.)
Seems to be a problem in the code.

E-mail/pm will be sent.

[James Heinrich]

I think I see what you're getting at. If you copy-paste the exponent with other characters in it (e.g. "M79,777,211") into the exponent field it would first limit the input to 9 characters, then remove the non-numeric characters, leaving you with a truncated value. I have removed the input length limit so it should work as expected now.

[Uncwilly]

[axn]

https://www.mersenne.ca/exponent/3243941 has an invalid factor that is messing up here and here

[James Heinrich]

Originally Posted by axn

https://www.mersenne.ca/exponent/3243941 has an invalid factor that is messing up here and here

Thanks, I have removed the invalid factor record. I have also fixed the code that was supposed to notify me about it but didn't.