What does the message after a test is done mean?

[ny1050220]

I have just finished the primality test for M51873683.
The message reads "M51873683 is not prime. Res64:0DD7C565A7CD80__. Wg4: 685C17CA,13956248,00000000".
The first sentence is in plain language, but what do Res64 and Wg4 mean?

[Prime95]

Congratulations on completing your first test!

Res64 is shorthand for "64-bit residue". This is the low 64-bit of the final 50 million bit result. We compare these 64 bits to the previous primality test and if they match we are certain that no hardware errors occurred in the two tests and the number is surely composite. You may also have seen a message like "successfully completes double-check..." indicating the 64-bits matched.

Wg4 is just a code indicating the which program version you used. Some of the other numbers are a checksum, a "shift count", and number of errors detected during the test.

You are now ready to start doing first-time primality tests. Best of luck!

[kriesel]

Originally Posted by ny1050220

I have just finished the primality test for M51873683.
The message reads "M51873683 is not prime. Res64:0DD7C565A7CD80__. Wg4: 685C17CA,13956248,00000000".
The first sentence is in plain language, but what do Res64 and Wg4 mean?

That's a prime95 primality test result record. (Maybe your first, a double check?)

Res64 is the least 64 bits of the final result of the primality computation (residue), after taking the pseudorandom shift (offset) off, for comparison to another run to see if they got the same result (double check).
Wg4 is a code for what version of prime95 was used.
I forget what the next two fields are, maybe security code and shift, but you appear to be in luck on the last one, all zeroes; that's a collection of error counts of differing types into one 32-bit word, also expressed in hexadecimal.
Yup, 13956248 was the shift value. See
https://www.mersenne.org/report_expo...exp_hi=&full=1

Generally it's bad form to post a full residue, for any unmatched test. Until there's a match, the right byte (2 hex chars) get masked on the server. This one is ok because it's a matching double check.

And this result puts to rest an exponent with 8 previous abandoned double checks, so bravo on that.

[ny1050220]

Thanks to you both for answering!

[ny1050220]

More questions about the results...
I've finished the test for M90218477, not a prime. The result looks like a 16 digit hexadecimal number, but the last two digits appear as __. Are the last two digits masked by the underscore so that when another CPU verifies the results, this is blind to it and it won't cheat?

[Uncwilly]

That is right. We don't want anyone to falsely submit a result. Sometimes tests go bad and we need 2 matching residues. By hiding the last 2 it makes it unlikely that someone would submit a matching fake result (and makes the chance of it matching an actual bad test even lower.)

[ny1050220]

Thanks!

[Chuck]

Originally Posted by ny1050220

More questions about the results...
I've finished the test for M90218477, not a prime. The result looks like a 16 digit hexadecimal number, but the last two digits appear as __. Are the last two digits masked by the underscore so that when another CPU verifies the results, this is blind to it and it won't cheat?

It looks like you are trying to double-check your own work. That needs to be done by a different user.