>After running the MMX Windows port for 12 hours on an
>AMD K6-2 450mhz, the program terminated with this error:
>*** bad distinguished point detected! ***
The program does a self-check from time to time before saving the
state. If the check fails it prints that message and quits without
saving since something has gone seriously wrong.
The most likely causes are memory or registers being corrupted
somehow. Maybe by a cosmic ray, a power fluctuation, a buggy kernel
or perhaps flakey hardware.
>4 points had already been detected and submitted by
>HTTP. I have since restarted the program and it is
>running smoothly again. Should I be worried?
When you restart it reads the last state saved before the self-check
failed. It then follows the exact same calculations except that there
will presumably be no corruption the second time around. It should
work as normal saving state from time to time and finding
distinguished points.
If by any chance it fails the self-check again before saving state or
finding a distinguished point, then there could be a bug in the binary
executable file you are running. In that case please contact me with
details about your set-up and send a copy of the saved state file.
Bye,
Rob.
This archive was generated by hypermail 2b29 : Sat Jan 01 2000 - 15:26:58 MET