 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hello,
I tried to render df3 volume file with 2048x2048x512 data and
povray crushes. the reason was pointer index for map.It was out from integer
range in the df3 reading routine in the file:
pattern.cpp
/////////////
line 4288: void Read_Density_File(DENSITY_FILE *df)
{
// int x, y, z, sx, sy, sz;
long x, y, z, sx, sy, sz;//changed int to long
///////////////
This works for reading of df3 file, but after parsing it is crushing
/////////
0:00:00 Rendering line 1 of 240Segmentation fault
/////////////////////
Where I need to change (x, y, z, sx, sy, sz) to get the run for huge data?
Thanks
Arman.
+++++++++++++++++++++++++++++++++
Arman Khalatyan
Address:
Astrophysical Institute Potsdam,
An der Sternwarte 16,
D-14482 Potsdam, GERMANY
Phones: (+049) 0331 7499517(work)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
arm2arm <nomail@nomail> wrote:
> I tried to render df3 volume file with 2048x2048x512 data
That amount exactly hits the 2GB barrier. Are you using the 64-bit
version of POV-Ray (which naturally requires you to have a 64-bit
system) or the 32-bit version?
If you are using the 64-bit version of POV-Ray then it is very possible
that there may still be some non-64-bit-compatible code left in POV-Ray
regarding this. However, it's important to know your system specs before
any further investigation.
> and povray crushes.
Btw, it's "crashes".
--
- Warp
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Sorry, the system is:
64bit
POV-Ray 3.6.1 compiled with PGI compiler and g++ 4.0.2 20050901 (Both are
segmenting )
RAM 32G
4CPU SMP maschine with AMD Opteron(tm) Processors 850.
If array size is 2048*2048*512 it is little bit bigger than Max_INT 2^31-1.
Therefore when we trying to read with int index in the file pattern.cpp the
index for array offset is calculated wrong.
also in the line when we try to check file type of df3 it is always telling
wrong file type due to the line/s: 4315
// figure out the data size
if((sx * sy *sz*4) == len)// this is always false if file size is bigger
4G
arman.
+++++++++++++++++++++++++++++++++
Arman Khalatyan
Address:
Astrophysical Institute Potsdam,
An der Sternwarte 16,
D-14482 Potsdam, GERMANY
Phones: (+049) 0331 7499517(work)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
arm2arm <nomail@nomail> wrote:
> Sorry, the system is:
> 64bit
I notified this issue to the povteam. It may indeed be that that part
of the code has not been fixed for 64-bit compatibility.
--
- Warp
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
arm2arm wrote:
> Hello,
> I tried to render df3 volume file with 2048x2048x512 data and
> povray crushes. the reason was pointer index for map.It was out from integer
> range in the df3 reading routine in the file:
> pattern.cpp
> /////////////
> line 4288: void Read_Density_File(DENSITY_FILE *df)
> {
> // int x, y, z, sx, sy, sz;
> long x, y, z, sx, sy, sz;//changed int to long
> ///////////////
>
> This works for reading of df3 file, but after parsing it is crushing
> /////////
> 0:00:00 Rendering line 1 of 240Segmentation fault
> /////////////////////
>
> Where I need to change (x, y, z, sx, sy, sz) to get the run for huge data?
> Thanks
The density file pattern is apparently indeed currently not suited for
being used above the 2GB limit. This is the case both in the DF3
reading you pointed out and the actual pattern evaluation. The required
changes are probably not very difficult but only recommended to make for
someone who knows the POV-Ray source code.
Testing this kind of thing is a bit cumbersome of course - having 2GB+
df3s lying around and reading them...
Christoph
--
POV-Ray tutorials, include files, Landscape of the week:
http://www.imagico.de/ (Last updated 20 Aug. 2006)
MegaPOV with mechanics simulation: http://megapov.inetart.net/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I made a patch for MegaPOV probably solving it (for 64bit systems only
of course):
http://megapov.imagico.de/cgi-bin/hgweb.cgi?cs=649a35101349
Note i tested it only with the standard 'galaxy.pov' - not with a large
file. Please check if it fixes your problem (the changes can also
easily be integrated into plain POV-Ray 3.6.1 manually if you don't want
to install Megapov).
Christoph
--
POV-Ray tutorials, include files, Landscape of the week:
http://www.imagico.de/ (Last updated 20 Aug. 2006)
MegaPOV with mechanics simulation: http://megapov.inetart.net/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi Christoph,
thank you for patch,
I never used megapov, can you provide the link for source code of megapov?
Thanks.
Arman.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> I never used megapov, can you provide the link for source code of megapov?
http://megapov.inetart.net/download.html
Here are some instructions to apply the patch to the megapov for Unix
source code:
1) Download the "generic Unix source version", for instance:
http://megapov.inetart.net/packages/unix/megapov-1.2.1.tar.bz2
2) Unpack it (creates the megapov-1.2.1/ directory):
tar xvjf megapov-1.2.1.tar.bz2
3) View Christoph's patch in raw form, and save it as a text file in your
home directory, e.g. ~/mp.patch:
http://megapov.imagico.de/cgi-bin/hgweb.cgi?cs=649a35101349;style=raw
4) Go to the megapov-1.2.1/ source directory and apply the patch:
cd megapov-1.2.1/
patch -p1 < ~/mp.patch
You should get something like this:
patching file source/frame.h
patching file source/patches/patches.h
Hunk #1 succeeded at 187 with fuzz 2 (offset -7 lines).
patching file source/pattern.cpp
Hunk #1 succeeded at 1739 (offset -279 lines).
Hunk #2 succeeded at 1761 (offset -279 lines).
Hunk #3 succeeded at 1780 (offset -279 lines).
Hunk #4 succeeded at 1807 (offset -279 lines).
Hunk #5 succeeded at 1871 (offset -279 lines).
Hunk #6 succeeded at 4607 (offset -332 lines).
Hunk #7 succeeded at 4634 (offset -332 lines).
Hunk #8 succeeded at 4651 (offset -332 lines).
Hunk #9 succeeded at 4668 (offset -332 lines).
5) Run the usual ./configure COMPILED_BY=your_name && make
- NC
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
COOL!
Everything is working for me,
I rendered 8Gb data.
Only the bottleneck is reading of df3 file,
I changed the zyx loops to z loop over xy slabs.
Thank you again.
Arman.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
arm2arm <web.44f5c4822dd4e7f6b399450c0@news.povray.org> Wednesday 30 of
August 2006 19:01
> COOL!
> Everything is working for me,
> I rendered 8Gb data.
> Only the bottleneck is reading of df3 file,
> I changed the zyx loops to z loop over xy slabs.
> Thank you again.
Interesting, can You share the patch that does it?
--
> wiesz. Ja wiem co to kernel. I wiem sporo o linuksie.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
