[NCLUG] not-lame segfaulting & dumping core

[S. Luke Jones]

Beginning summer 1999 I encoded a zillion tracks using notlame
(also blade, but that's beside the point) and recall only having
a couple of core dumps.

A couple of months ago I freed up some memory and upgraded my
box from 64 MB to 128 MB. About the same time I began having a
core dump from not-lame at least once per album (probably about
1 per 40 min of .wav file).

Normally I tend not to blame the hardware (I blame the programmer
unless it was me :-) but this time I'm suspicious due to the
correlation in time between new (old) RAM and increased dumping
of core.

Yes, I can go pulling the RAM out and try before/after etc. but
the motherboard is a Tyan 1572 and I'm tossing it out and buying
something built in the ATX era the next time I open up the box,
much less try to get my chubby fingers into the spot where the
!#!@%!& SIMM sockets are. So let's assume hardware swapping is
not an option.

Are there any good user space RAM testers out there? How about
kernel space? How about purified versions of notlame that won't
segfault? Etc.

How would y'all trouble-shoot this one?

-- 
Luke Jones = luke/vortex/frii/fullstop/com