Duplicate lines in HRF3.TXT

[GP2]

There are some fully-duplicated lines in HRF3.TXT:


16073131,telos,CC304EA96,WX5

occurs 9 times (there are no other results for this exponent).


12136727,Northwestern,CCEDC533B,WW1

occurs 7 times (there are no other results for this exponent).


11947847,curtisc,ccc61,WV1

occurs 3 times (there are no other results for this exponent)


10916119,nestor,Athlon,WW1

occurs 3 times, and there are 2 other results by GaBi/Athlon (the same machine?) and by S107595/C25D19D06

And 49 other cases where the same line occurs 2 times.

Did telos/CC304EA96 really return 9 different residues for the same exponent?

By the way, even assuming that there are 8 errors out of those 9 results, telos/CC304EA96 still wouldn't qualify as an error-prone machine: 0 bad results, 40 good results, 465 unverified needing a second test, 26 unverified needing a third or higher test.

For Northwestern/CCEDC533B, it would qualify as an error-prone machine because we must assume there are 6 errors out of the 7 results. However this machine has never returned results for any other exponents and has no assignments either.

[Prime95]

Originally posted by GP2
Did telos/CC304EA96 really return 9 different residues for the same exponent?

Yes.

[GP2]

Originally posted by Prime95
Yes.

Whoa.

How could that happen? And why were they even re-testing it 9 times in the first place?

I'll propose this for the Mersenne-aries forum... an interesting mystery.