 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Thanks Andrew.
I made a fresh povray-3.50c-bsp/ directory and copied your src/ into
povray-3.50c-bsp/src
However, after ./configure the make failed with the same error as before:
"your implementation of AM_INIT_AUTOMAKE comes from an older version of
automake" and to run aclocal
After running aclocal and trying again I still got that whole bunch of
errors where it cant find some mysterious ".Po" files (not a typo). That's
from the Makefile within povray-3.50-bsp/src/ lines 256 thru 352.
Doing "make --version" shows GNU make 3.79.1
Anything else you need to know?
Thanks,
George
> Can you try installing the full set of modified files:
>
> http://povplace.addr.com/bsppov-src.tar.gz
>
> and let me know if you still have problems compiling? Make sure that you
> start with a completely fresh povray-3.50c.
>
> Andrew
>
>
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
>I made a fresh povray-3.50c-bsp/ directory and copied your src/ into
>povray-3.50c-bsp/src
>
>However, after ./configure the make failed with the same error as before:
>"your implementation of AM_INIT_AUTOMAKE comes from an older version of
>automake" and to run aclocal
>
>After running aclocal and trying again I still got that whole bunch of
>errors where it cant find some mysterious ".Po" files (not a typo). That's
>from the Makefile within povray-3.50-bsp/src/ lines 256 thru 352.
>
>Doing "make --version" shows GNU make 3.79.1
>
>Anything else you need to know?
>
>Thanks,
>George
>
Can you successfully build povray before overwriting with patched files?
This should at least make sure you have a working automake... If you can't
build the source without the patched files you probably won't be able to
build it with the patch.
Andrew
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I can indeed build povray-3.50c, Andrew.
I just tried copying all the modified source files, but not any of the
Makefiles, and I came very close to compiling it.. but it errored out with
"undefined reference to Instersect_Bsp_Tree()" and several other BSP
related errors like that.
Besides adding bsptree.cpp to the list of source files in the Makefile,
what else would I have to add to it?
Thanks,
George
>>
>
> Can you successfully build povray before overwriting with patched files?
> This should at least make sure you have a working automake... If you
> can't
> build the source without the patched files you probably won't be able to
> build it with the patch.
>
> Andrew
>
>
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hello again,
I've compiled a version of povray3.5c with bsp-tree patch on a Fedora 1
linux (gcc 3.3.2, and only added -mcpu=athlon-xp to the options). Tested
with benchmark.pov. Not much differences in rendering time (slighty faster
with BSP tree) but there is a change in the image , a small part at the
left-bottom of the checkered vase have disappeared in the povray-bsp version
M
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Mael wrote:
>Hello again,
>
>I've compiled a version of povray3.5c with bsp-tree patch on a Fedora 1
>linux (gcc 3.3.2, and only added -mcpu=athlon-xp to the options). Tested
>with benchmark.pov. Not much differences in rendering time (slighty faster
>with BSP tree) but there is a change in the image , a small part at the
>left-bottom of the checkered vase have disappeared in the povray-bsp version
>
>M
>
Hmmm, I tried rendering benchmark.pov yesterday at a low resolution and
didn't see any difference. I'll try it again today a bit larger.
About the performance of this scene (and in general), benchmark.pov only
contains 157 objects (as I remember) which isn't really enough to exercise
the bounding code too much. Also, it uses a bunch of other advanced
features that will eat away the rendering time. When I get the patch to
work with some more object types, it may improve some more but I would not
expect spectacular improvements on an arbitrary scene. The 70% improvement
was on an extremely isolated case.
It might be worth it to try on some IRTC scenes that have large numbers of
objects. At this point I've only had the patience to try it on a few,
because most of the scenes I've downloaded don't render out of the box.
Perhaps some others who have successfully compiled it could give this a
try? :)
Andrew
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
George Pantazopoulos wrote:
>I can indeed build povray-3.50c, Andrew.
>
> I just tried copying all the modified source files, but not any of the
>Makefiles, and I came very close to compiling it.. but it errored out with
>"undefined reference to Instersect_Bsp_Tree()" and several other BSP
>related errors like that.
>
>Besides adding bsptree.cpp to the list of source files in the Makefile,
>what else would I have to add to it?
>
>Thanks,
>George
>
Can you try one more thing? Start from clean 3.50c source again, the try
using the updated files I've posted:
http://povplace.addr.com/bsppov-src.tar.gz
All I have done is make the source file list alphabetical in Makefile.am.
I've noticed some weird things in the generated Makefile for my original
Makefile.am, that it seems my version of make was ignoring. I hope that
this works, if you still have problems let me know. I don't really know
why the list needs to be alphabetical. If this still doesn't work, try
adding bsptree.* after every occurance of bsphere.* in the generated
Makefile and see if that works. If you still can't do it, I'm not sure
what to suggest :(
Andrew
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Hmmm, I tried rendering benchmark.pov yesterday at a low resolution and
> didn't see any difference. I'll try it again today a bit larger.
Maybe it's a problem with my compilation.. ? You can quickly check by
deactivating photon and area light then run
povray +Ibenchmark.pov -w800 -h800 +sc0.043716 +sr0.273224 +ec0.153005
+er0.371585
povraybsp +Ibenchmark.pov -w800 -h800 +sc0.043716 +sr0.273224 +ec0.153005
+er0.371585 +MM2
I can post the images in one of binaries newgroups if you want
M
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Andrew Clinton wrote:
> I've put together the patch as a set of diffs (following the example of
> the
> previous patch posted here). The patch can be found at
> http://povplace.addr.com/bsppov.tar.gz
>
Patched my unix sources without trouble.
Test scene with about 120 CSG objects (only torus/box/cylinder and the
like) rendered about 1 percent slower, however, while comsuming some
kb more memory.
I have little clue of the details of your patch but
maybe you could consider retaining the vista buffer because it is said
to be cheap & fast.
Just my 2 cents...
Wolfgang
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Mael wrote:
>> Hmmm, I tried rendering benchmark.pov yesterday at a low resolution and
>> didn't see any difference. I'll try it again today a bit larger.
>
>Maybe it's a problem with my compilation.. ? You can quickly check by
>deactivating photon and area light then run
>povray +Ibenchmark.pov -w800 -h800 +sc0.043716 +sr0.273224 +ec0.153005
>+er0.371585
>povraybsp +Ibenchmark.pov -w800 -h800 +sc0.043716 +sr0.273224 +ec0.153005
>+er0.371585 +MM2
>
>I can post the images in one of binaries newgroups if you want
>
>M
>
You are right, there is a bug in the pruning system. I'll fix this...
Andrew
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Wolfgang Wieser wrote:
>
>Test scene with about 120 CSG objects (only torus/box/cylinder and the
>like) rendered about 1 percent slower, however, while comsuming some
>kb more memory.
>
CSG seems poorly bounded by default. Are you rotating or scaling the CSG?
If so, try a scene without these transformations and see if there is an
improvement. Also try it with more objects (>1000).
>I have little clue of the details of your patch but
>maybe you could consider retaining the vista buffer because it is said
>to be cheap & fast.
>
Light/Vista buffers are too tightly integrated with the bounding box
hierarchy to be easily adapted for use with the BSP tree, at least for now.
I'll need to give some more thought to how a screen space hierarchy could
benefit with the tree. There are other things that will improve
performance more which I am going to try first (such as construct better
bounding boxes for csg).
Andrew
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
