"SharkD" <nomail@nomail> wrote:
> "Roman" <nomail@nomail> wrote:
> > Hi everyone
> >
> > this is my first contribution here.
> > I'm running the 64bit beta 29 on Windows Vista x64. Intel quadcore CPU, 4GB RAM.
> >
> > I'm trying to render a rather long sequence of images with radiosity with the
> > built-in queue feature. The first scene renders problem-freely, but the
> > subsequent ones in the queue fail with an "unknown internal error".
>
> I've been experiencing the same error.
>
>
http://news.povray.org/povray.bugreports/thread/%3Cweb.493eda204f59c5434f5a8c2a0%40news.povray.org%3E/
>
> However, I wasn't as thorough about trying to identify the causes. I didn't
> think of trying to see what would happen if I disabled radiosity.
>
> -Mike

Exactly! I get the "Parse Error: Cannot open input file", too. Did you also use
the 64bit version? If not, maybe using the 32bit exe would be a workaround for
the moment...

I think I might have to set up a minimal example scene which produces the error
so the devs can have a look at it.

In the meantime: Nobody know a place where I could get beta 15?

Post a reply to this message

"Roman" <nomail@nomail> wrote:
> Exactly! I get the "Parse Error: Cannot open input file", too. Did you also use
> the 64bit version? If not, maybe using the 32bit exe would be a workaround for
> the moment...
>
> I think I might have to set up a minimal example scene which produces the error
> so the devs can have a look at it.
>
> In the meantime: Nobody know a place where I could get beta 15?

No, I've been using the 32-bit version. It's annoying, because it makes
rendering animations a real pain in the butt.

-Mike

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> "Roman" <nomail@nomail> wrote:
> Hm... having embarked on the quest of fixing some issues in the 3.7 radiosity, I
> must confess I'm less interested in solving your immediate problems, but rather
> in learning what's wrong with the .29 radiosity that might cause the crash in
> the first place...
>
> So can you narrow down the crash to a particular scene (or maybe even particular
> aspects of it), and post that one in one of the binaries newsgroups, so I can
> check that?
>
> > Any help ist highly appreciated!
>
> Same goes for me :)

It appears to me that reducing the radiosity count from 300 to 200 solves the
problem, but that's of course no satisfactory solution. I added this
information to
http://news.povray.org/povray.bugreports/thread/%3Cweb.493eda204f59c5434f5a8c2a0%40news.povray.org%3E

Post a reply to this message

"Roman" <nomail@nomail> wrote:
> I think I might have to set up a minimal example scene which produces the error
> so the devs can have a look at it.

Oh, yes please! (*rubs his hands in excitement for another scene to test on*)

> In the meantime: Nobody know a place where I could get beta 15?

I'd be surprised if there was. I guess officially it wouldn't be allowed to
distribute it anymore - let alone that I'd expect it to have lots of other
drawbacks.

Post a reply to this message

"Roman" <nomail@nomail> wrote:
> It appears to me that reducing the radiosity count from 300 to 200 solves the
> problem, but that's of course no satisfactory solution. I added this
> information to
>
http://news.povray.org/povray.bugreports/thread/%3Cweb.493eda204f59c5434f5a8c2a0%40news.povray.org%3E

I didn't notice until now that we're talking about 64 bit. But I'll try to see
if it also happens on 32-bit, or 64-bit Linux.

At the moment I suspect that it's actually all about timing, and that the 3.7
beta .29 may have a general problem with accessing the same file from multiple
threads at the same time - although I actually can't think of any file that
would be opened by both renders of that scene.

Do you have the source files still open in POV when you start the render?

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> Do you have the source files still open in POV when you start the render?

No, I don't.

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> I didn't notice until now that we're talking about 64 bit. But I'll try to see
> if it also happens on 32-bit, or 64-bit Linux.

As SharkD mentioned, the error also occurs in the 32bit version:
http://news.povray.org/povray.beta-test/message/%3Cweb.4957638b86257958d2ca984a0%40news.povray.org%3E/#%3Cweb.4957638b8
6257958d2ca984a0%40news.povray.org%3E
I myself haven't checked, though.


> At the moment I suspect that it's actually all about timing, and that the 3.7
> beta .29 may have a general problem with accessing the same file from multiple
> threads at the same time - although I actually can't think of any file that
> would be opened by both renders of that scene.

But then why does it work with a radiosity count of less than ~300? To me, from
an outside point of view, it seems more likely that there is some memory
problem with the radiosity feature (or somewhere else). It looks as if some
allocated memory wasn't freed properly, causing trouble when POV-Ray wants to
go on with the next scene file, or that an array overflew. Also the memory dump
error popup (saying: unhandled exception) in the end of rendering a single scene
gives rise to such an assumption.

Post a reply to this message

"Roman" <nomail@nomail> wrote:
> As SharkD mentioned, the error also occurs in the 32bit version:
>
http://news.povray.org/povray.beta-test/message/%3Cweb.4957638b86257958d2ca984a0%40news.povray.org%3E/#%3Cweb.4957638
b8
> 6257958d2ca984a0%40news.povray.org%3E

Not seen here yet. (Not when leaving the system alone while it's rendering, at
least; I've seen some errors regarding file access stuff, but usually I was
loading something else in the editor at that very same moment.)

> > At the moment I suspect that it's actually all about timing,

> But then why does it work with a radiosity count of less than ~300?

Maybe because it just changes the timing.

> To me, from
> an outside point of view, it seems more likely that there is some memory
> problem with the radiosity feature (or somewhere else).

How much RAM do you have available? I'm running on 2 GB here, and don't see any
such problems, although I'm running other stuff in parallel, like MS Visual
Studio and the like.

> It looks as if some
> allocated memory wasn't freed properly, causing trouble when POV-Ray wants to
> go on with the next scene file, or that an array overflew. Also the memory dump
> error popup (saying: unhandled exception) in the end of rendering a single scene
> gives rise to such an assumption.

"Unhandled exception" could be just about anything.

Could also be related to thread synchronization. Maybe some cleanup stuff not
fully multithread-safe, and two threads trying to clean up at the same moment.

Unfortunately it's very hard to pin down such issues unless they can be
reproduced on a system with a debugger running.

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> How much RAM do you have available? I'm running on 2 GB here, and don't see any
> such problems, although I'm running other stuff in parallel, like MS Visual
> Studio and the like.

I have 4GB, of which over 2.5GB are free ("cached" in the task monitor). Memory
shouldn't be a problem.

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> "Roman" <nomail@nomail> wrote:
> > As SharkD mentioned, the error also occurs in the 32bit version:
> >
http://news.povray.org/povray.beta-test/message/%3Cweb.4957638b86257958d2ca984a0%40news.povray.org%3E/#%3Cweb.49576
38
> b8
> > 6257958d2ca984a0%40news.povray.org%3E
>
> Not seen here yet. (Not when leaving the system alone while it's rendering, at
> least; I've seen some errors regarding file access stuff, but usually I was
> loading something else in the editor at that very same moment.)

I've just tried the 32bit version. Same problem, but instead of the "parse
error: Cannot open input file.", the error now reads "Unknown internal error!".

Post a reply to this message