#Pharlap error 10118
14 messages in this thread
Thanks for the reply, Alan, but that would be a major task as I am using
special formatting for this ctrl. card for each of 4 IDE drives. It never has
caused any problems before. I did a Sdefrag on the stacker drive where 3DS
resides and finished the animation project with further incident. I could
understand that bugs in pharlap getting hung on a cache ctrl. but to corrupt
the 3DS.exe file? That is really puzzling.
3DS residing on a stacker drive? I've definitely heard of executables becoming
corrupt when stacked, although I've never seen it myself. Maybe that was your
problem…
-Alan
I;ve been running 3DS on a stacked drive for over two years with absoultely no
problems. I use Stacker though. It seems to be a very stable product. Just
thought I'd let you know.
Dave Mennenoh
Hey David, a fellow stacker lover.
I've got my eye on a cache controller card from PSI that puts stacker on the
controller. A daughter board just plugs into the main controller. PSI is a
licensed mfg. of Stacker technology. Advantages: no drivers to load and
everything is transparent to you when using this board. Once I figure out what
my next MB is going to be either EISA/VLB or VLB/ISA. I will choose which
version of this board I will get for the HD. I've been using an IDE ISA cache
ctrl. for two years now and am very pleased with its performance: 3.3 MB/sec
sustained transfer rate. Disadvantages are these type boards are operating
system specific and so is stacker. If I go to OS/2 or win NT I would be SOL.
Don,
Cool idea – putting stacker right on the controller. I'll bet it will be much
faster – and not having to load the drivers would be great. BTW, I know STAC is
making an OS/2 specific version of stacker – if that's worth anything.
Dave Mennenoh
Believe it or not, things like that (weird corrupting of the 3DS.EXE file) can
and do happen. Why, when, how, where, but I've seen it personally and in
numerous messages here.
Of course, I've always believed our office was built over an ancient Indian
burial ground!<g>
(No offense to any Native Americans intended or implied)
Dave
Once you start swapping, Phar Lap will increase the size of your swap file for
each frame… it's just the way PL works… it swaps extremely inefficiently.
Basically, you've got to free up much more disk space or get enough RAM to keep
from swapping.
– G
Gary: I think that was what my problem was because I moved the swap dir to a
stacker drive and it rendered OK, I thought I'd heard you couldn't do this so
I just kept the swap dir on an unstacked drive but I had much less space there.
Anyway, is that what the error message means? That I didn't have enough
swapdir space? Why did my 3ds.exe get corrupted?
It sounds like that was the problem… that error message from PL supposedly
means that it can't allocate any more virtual memory. It probably corrupted
your 3DS.EXE because it tried to write a swap file over it.
– G
Gary,
A while back, I had a problem with a 3DS file where some of my objects mapping
coordinates were wiped out, but the material was still assigned. I found the
problem and it is swap file related.
If you begin swapping to disk and try to create/object/attach, you will get an
error saying there isn't enough RAM to attach, followed by some vertice count.
That sequence removes the mapping coordinates of the main object.
Obviously related to the Phar Lap memory swapping, FWIW.
Steve Ormonde
P.S. It was nice seeing you again at Siggraph, congrats on R3!
This Phar Lap stuff never ends… even though we probably can't do anything
about it, I'll put your comments on my "issues" list and will spend a little
time with it.
Thanks,
– G
Steve-
PMJI but I have a .prj file that hangs up at the "Assigning maps…"
part of the rendering. It souds like the same problem you encountered.
Did you find a solution to this or did you just have to reassign all
the mapping coordinates.
Thanks
Carl
Carl,
I was notified of my problem by a "Object01 needs mapping coordinates" prior to
getting to the "preparing maps" stage. The only fix is to reassign the mapping
coord.
This seems to happen after I've done a number of renderings in a single
session. I only have 16mb of RAM and when it starts swapping, that's when the
error occurs. To avoid it, I click "hold" prior to "object/attach" when I know
the swap file is large.
Steve Ormonde
Gary: thanks for the confirmation.
Wish List:– Why can't we either do away with error numbers and replace with
short explanations if it is so difficult to supply a *.txt file with these
message explanations along with the software?