Greetings.

I have a scene which contains some external JPEG graphic files. My aatempts 
to render the scene fail, due a lack of RAM (I only have 1.25 GB).

My idea is, and I want to place it here as a feature request, that POV-Ray 
should not only abort the parsing with an error message such as 
"Insufficient memory to allocate intermediate image storage", but to support 
the user more actively.

My wish is, that POV-Ray tells the user how many MB all images together 
should have in order to have the scene still being able to be rendered. That 
would help the user to downsize the graphic(s) accordingly to the size 
allowed as maximum, and prevents these endless try-and-run-POV-Ray until, 
somewhen when the sun sets down and the birds stop singing their endless 
pink songs of love and the fox says Good Night to the chicken, it finally 
works.

Is that possible?

Greetings,

Sven

Post a reply to this message

Sven Littkowski wrote:
> Greetings.
> 
> I have a scene which contains some external JPEG graphic files. My aatempts 
> to render the scene fail, due a lack of RAM (I only have 1.25 GB).

If POV-Ray fails due to lack of memory, it is due to *hardware* limitations.
Operating systems for 32 bit hardware make 2-3 GB of the maximum 4 GB
physically addressable available to a user program. If your scene does not
fit, you cannot render it, and there is nothing POV-Ray can do about it.

Further, POV-Ray cannot tell you how much memory it would need because it
never actually can parse the whole scene. You are asking for information
POV-Ray cannot give you without parsing the whole scene, which it cannot
exactly because you do have insufficient memory. As such, it is simply a
paradox to request POV-Ray to provide you with this information.

	Thorsten

Post a reply to this message

Thorsten Froehlich <tho### [at] trfde> wrote:
> Sven Littkowski wrote:
> > Greetings.
> >
> > I have a scene which contains some external JPEG graphic files. My aatempts
> > to render the scene fail, due a lack of RAM (I only have 1.25 GB).
>
> If POV-Ray fails due to lack of memory, it is due to *hardware* limitations.
> Operating systems for 32 bit hardware make 2-3 GB of the maximum 4 GB
> physically addressable available to a user program. If your scene does not
> fit, you cannot render it, and there is nothing POV-Ray can do about it.
>
> Further, POV-Ray cannot tell you how much memory it would need because it
> never actually can parse the whole scene. You are asking for information
> POV-Ray cannot give you without parsing the whole scene, which it cannot
> exactly because you do have insufficient memory. As such, it is simply a
> paradox to request POV-Ray to provide you with this information.
>
>  Thorsten

PovRay won't render a scene when you have 1.5GB? Wow, that's an achievement.
Anyway, there should still be a way as long as the parsing stage doesn't
deplete your memory, and that's by splitting the image into several pieces
and rendering them separately. I don't have the documentation in front of
me so I can't tell you the command-line arguments/ini flags, but there's
something to select x- and y-values to draw, and to continue a render that
wasn't completed.

Of course, if your scene is such that the jpegs + all the other
memory-intensive stuff (huge meshes come to mind) alone take up 1.5GB,
you'll just have to get more memory. Turning off the image-display might
also help, but don't quote me on that.

--Andrea Reina

Post a reply to this message