Hi Julian,

> Hi.  I'm an author of valgrind, a memory debugger which I believe at
> least one person has used on FlightGear.

We were three, AFAIK (Cameron, Andy and I). While I was quite active
at first, I had to stop when my tdfx-card got busted. Now I've got a
new machine with one of those nasty nVidia cards (I'm innocent ... almost!),
but when I tried to start valgrinding again recently (v1.9.4), I ran into
valgrind's lack of SSE support. (Wasn't keen to link against mesa. :-)
I'm a regular visitor at http://developer.kde.org/~sewardj/. Downloaded
valgrind from sourceforge now.



> We've just extended valgrind to handle MMX instructions, and it works
> on some simple-ish test cases.  I was wondering if FlightGear can be 
> configured to use MMX, and, if so, whether anyone would care to try
> running such a configuration on valgrind.

Unfortunately, my fgfs binary uses SSE (libGL.so.1.0.4191.nv_glx?) and
valgrind aborts with this message:

| disInstr: unhandled 2-byte opcode: 0x28 0x2 0xF
| This _might_ be the result of executing a SSE, SSE2 or 3DNow!
| instruction.  Valgrind does not currently support such instructions.  Sorry.
| Illegal instruction (core dumped)

But until that it seemed to work well.  :-)



> It would be useful for us, to gain early feedback on our MMX work.

What we can do for you is a joke, compared with what you already did
for us! Besides many sleeping bugs, there were some that caused crashes
and would have been a lot harder to find without valgrind! 



> We hope to find the time, soon, to attempt support of SSE/SSE2 instructions.

Sigh.   :-)

Thanks for that wonderful tool!
m.


_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel

Reply via email to