Wrong entries in database?

I just looked at the status of a few small exponents and I discovered sth. strange: Some (small) exponents seem to have wrong LL-Test marked as "Verified LL":

Here are three examples:

M10061
No factors below 2^59
P-1 B1=2147483647
Verified LL 3188 by "Unknown"
Verified LL 56EB9BB91825B188 by "George Woltman"
Verified LL 56EB9BB91825B188 by "Brian J. Beesley"

M10399
No factors below 2^58
P-1 B1=4290000000
Verified LL 151E by "Unknown"
Verified LL 7D445D0CFC69951E by "George Woltman"
Verified LL 7D445D0CFC69951E by "Brian J. Beesley"

M10667
No factors below 2^58
P-1 B1=1000000000
Verified LL 5B7B by "Unknown"
Verified LL AD23A344EA29DB7B by "George Woltman"
Verified LL AD23A344EA29DB7B by "Brian J. Beesley"
History 430 curves, B1=1000000, B2=100000000 by "ANONYMOUS" on 2007-09-27


Some 16bit LL-Residues submitted by "Unknown" seem to be wrong (M10061: "...151E" instead of th correct "...951E"). This is - as far as I'm understanding this - an error and should be corrected. I haven't the time to check more exponents, so I wrote it in the forum...

PS:
Here's the direct link I used:
http://v5www.mersenne.org/report_exp...&B1=Get+status

[robo_mojo]

Probably because the 16 bit residues are actually 15 bits (perhaps the top bit is a flag for something).

Ex (from your residues):

3 hex = 0011 binary
B hex = 1011 binary

1 hex = 0001 binary
9 hex = 1001 binary

5 hex =0101 binary
D hex =1101 binary

Just a guess.

[cheesehead]

I was just going to say the same: The four-hexadecimal-digit residues represent only 15-bit values.

Thank you anyway, "Unregistered", for your concern about this project and for taking the time and initiative to start this thread to ask your question !!! Your next question might lead to something that does need correction.

In fact, this one does point out a needed change, too: GIMPS documentation may not yet have an explanation, or as prominent a one as is needed, of the apparent discrepancy you noticed. Thank you.

[Prime95]

The 2 explanations are correct. I'm loathe to eliminate the small residues as these are in many cases, to the best of my knowledge, the very first LL test on the number.

[cheesehead]

... and the small residues shouldn't be eliminated. They are as genuine as the large ones.

- - -

George,

I need to put a definite date on my calendar (before the end of summer, say) for compiling my documentation suggestions and presenting proposed updated versions of various pages and text files to you. Please ask me about this if you don't see it by October.