ERROR: Factor doesn't divide N!

[chalsall]

I have five machines which spend about 50% of their time doing P-1 tests, so that other ones don't have to (and thus don't use more than ~15 MB of RAM). One of these (an AMD Duron @ 1300MHz) has 256 MB available for the work. When I checked the logs on this machine just now I noticed an error entry I hadn't seen before:

[code:1]
[Sat Apr 5 15:55:15 2003]
P-1 found a factor in stage #2, B1=215000, B2=3977500.
ERROR: Factor doesn't divide N!
UID: wabbit/wabbit, M18716707 completed P-1, B1=215000, B2=3977500, WY2: E4682B1A
[Sun Apr 6 12:56:11 2003]
P-1 found a factor in stage #1, B1=180000.
ERROR: Factor doesn't divide N!
UID: wabbit/wabbit, M15069997 completed P-1, B1=180000, B2=3825000, WY2: 85DB5A56
[Mon Apr 7 22:08:39 2003]
P-1 found a factor in stage #1, B1=215000.
ERROR: Factor doesn't divide N!
UID: wabbit/wabbit, M18717269 completed P-1, B1=215000, B2=3977500, WY2: E4662B01
[/code:1]
This is running version 22.12.2 under Linux.

Do I have a hardware problem, or is there something else going on here? Would it be worth re-running these P-1 tests under 23.2.2? This error has not appeared on the other P-1 machines, but they don't have as much RAM assigned to them.

[ET_]

Which version of the program are you using?

There was a small flaw in version 22.12. It affected machines with less than 75 MB of memory assigned to the task when doing P-1

You may try version 23.2

Luigi

[NickGlover]

Originally Posted by ET_

There was a small flaw in version 22.12. It affected machines with less than 75 MB of memory assigned to the task when doing P-1

This flaw was present in 22.11. It was fixed in 22.12.

[ET_]

Whoops... you're right!

22.12 corrected that bug. sorry :-} ops:

Luigi

[chalsall]

Just a follow up -- I reran the P-1 test under 23.2.2 for M15069997 (excact same memory settings), and did not get the same error. (Didn't find a factor either; alas.)

I'll probably try rerunning it again under 22.12.2 to see if it's reproducable.