Re: [Bf-committers] How to mix C and C++ in Blender
On Wed, Jun 19, 2013 at 7:29 AM, Jürgen Herrmann wrote:
> Hi Alexander,
>
> I personally don't like this, but it is possible:
>
> http://stackoverflow.com/questions/7281441/elegantly-call-c-from-c
> http://stackoverflow.com/questions/2399344/calling-c-code-from-c
> http://stackoverflow.com/questions/2744181/how-to-call-c-function-from-c
>
> Google is your friend ;) Sometimes...
>
> I don't know exactly what the blender coding conventions say about this, but
> I would consider this bad style, because you have to introduce wrappers that
> might easily break when a external lib is updated.
> I am curious what the other devs say ;)
>
> /Jürgen
>
> Am 18.06.2013 um 23:02 schrieb Alexander Pinzon Fernandez
> :
>
>> Hi all.
>>
>> I am working on Bmesh operator for perform laplacian editing for sketch
>> mesh gsoc project.
>>
>> If the library to solve the linear system is SuperLU and Eigen3 for matrix
>> operations, then I must learn how to mix C code (SuperLU) and C ++ (Eigen3)
>> in Blender. (elegant and organized way).
>>
>> I have c code file in
>> source\blender\bmesh\operators\bmo_deform_laplacian.c
>> in bmo_deform_laplacian.c i call several c++ functions (ej. au = ftest(4
>> ,5);) declare in bmo_deform_utils.h and implemented in
>> bmo_deform_utils.cpp.
>>
>> this two files are in source\blender\bmesh\operators\
>>
>> What are the directories where the files should put bmo_deform_utils.h
>> and bmo_deform_utils.cpp?
>> What are policies and style guides in Blender to mix C code and C ++?
>>
>> Example
>>
>> bmo_deform_laplacian.c
>> //
>>
>> #Include "bmo_deform_utils.h"
>> ...
>> au = ftest(4 ,5);
>> ...
>> ///
>>
>> bmo_deform_utils.h
>> ///
>> #ifndef __BMO_DEFORM_UTILS_H__
>> #define __BMO_DEFORM_UTILS_H__
>>
>>
>> #ifdef __cplusplus
>> extern "C" {
>> #endif
>>
>> int ftest(int x, int y);
>>
>> #ifdef __cplusplus
>> };
>> #endif
>>
>> #endif //__BMO_DEFORM_UTILS_H__
>> ///
>>
>> bmo_deform_utils.cpp
>> ///
>>
>> #include "bmo_deform_utils.h"
>>
>> class ABC{
>> public:
>> int a,b;
>> ABC(int ax, int by ){
>> a = ax;
>> b = by;
>> }
>> int fsum(){
>> return a + b;
>> }
>> };
>>
>> int ftest(int x, int y){
>> ABC A(x, y);
>> return A.fsum();
>> }
>> ///
>>
>> Thanks in advance.
>>
>> Att
>> Alexander Pinzon Fernandez
In blender we have a convention of _not_ mixing in C++ with existing C code.
For examples of C code that calls C++ in Blender see:
extern/rangetree/range_tree_c_api.h (used by BMesh)
extern/rangetree/range_tree_c_api.cc
intern/rigidbody/RBI_api.h
intern/rigidbody/rb_bullet_api.cpp
intern/audaspace/intern/AUD_C-API.h
intern/audaspace/intern/AUD_C-API.cpp
--
- Campbell
___
Bf-committers mailing list
Bf-committers@blender.org
http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Windows buildbot naming
That is indeed much better, thank you. :) Am 19.06.2013 00:18, schrieb Brecht Van Lommel: > I've made some changes, hopefully it's all clear now. > > On Tue, Jun 18, 2013 at 10:57 PM, Thomas Dinges wrote: >> Hi, >> a few people and I just had a discussion in IRC about buildbot being >> confusing. >> >> Imho the current state is a bit problematic, most people will just >> search for "Windows 64" and download a build. >> >> Could we please communicate better that the vc2008 is the official one still >> and vc2012 and multi view are experimental? Maybe add a label like for >> MinGW. >> >> Or maybe the web UI allows some grouping of "Official Trunk builds" and >> "Branches / Experimental"? >> >> Thanks, >> Thomas >> >> >> -- >> Thomas Dinges >> Blender Developer, Artist and Musician >> >> www.dingto.org >> >> ___ >> Bf-committers mailing list >> Bf-committers@blender.org >> http://lists.blender.org/mailman/listinfo/bf-committers > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers -- Thomas Dinges Blender Developer, Artist and Musician www.dingto.org ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Windows buildbot naming
I've made some changes, hopefully it's all clear now. On Tue, Jun 18, 2013 at 10:57 PM, Thomas Dinges wrote: > Hi, > a few people and I just had a discussion in IRC about buildbot being > confusing. > > Imho the current state is a bit problematic, most people will just > search for "Windows 64" and download a build. > > Could we please communicate better that the vc2008 is the official one still > and vc2012 and multi view are experimental? Maybe add a label like for > MinGW. > > Or maybe the web UI allows some grouping of "Official Trunk builds" and > "Branches / Experimental"? > > Thanks, > Thomas > > > -- > Thomas Dinges > Blender Developer, Artist and Musician > > www.dingto.org > > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Windows buildbot naming
Hi, That is a Great idea, I too had problems to find the right builds after the vs2012 x64 buildbot finished for the first time. We should also indicate that the vs2012 build won't run on win XP for now. I found a solution for this but it'll take some time to redeploy the libs involved. /Jürgen Am 18.06.2013 um 22:57 schrieb Thomas Dinges : > Hi, > a few people and I just had a discussion in IRC about buildbot being > confusing. > > Imho the current state is a bit problematic, most people will just > search for "Windows 64" and download a build. > > Could we please communicate better that the vc2008 is the official one still > and vc2012 and multi view are experimental? Maybe add a label like for > MinGW. > > Or maybe the web UI allows some grouping of "Official Trunk builds" and > "Branches / Experimental"? > > Thanks, > Thomas > > > -- > Thomas Dinges > Blender Developer, Artist and Musician > > www.dingto.org > > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] How to mix C and C++ in Blender
Hi Alexander,
I personally don't like this, but it is possible:
http://stackoverflow.com/questions/7281441/elegantly-call-c-from-c
http://stackoverflow.com/questions/2399344/calling-c-code-from-c
http://stackoverflow.com/questions/2744181/how-to-call-c-function-from-c
Google is your friend ;) Sometimes...
I don't know exactly what the blender coding conventions say about this, but I
would consider this bad style, because you have to introduce wrappers that
might easily break when a external lib is updated.
I am curious what the other devs say ;)
/Jürgen
Am 18.06.2013 um 23:02 schrieb Alexander Pinzon Fernandez :
> Hi all.
>
> I am working on Bmesh operator for perform laplacian editing for sketch
> mesh gsoc project.
>
> If the library to solve the linear system is SuperLU and Eigen3 for matrix
> operations, then I must learn how to mix C code (SuperLU) and C ++ (Eigen3)
> in Blender. (elegant and organized way).
>
> I have c code file in
> source\blender\bmesh\operators\bmo_deform_laplacian.c
> in bmo_deform_laplacian.c i call several c++ functions (ej. au = ftest(4
> ,5);) declare in bmo_deform_utils.h and implemented in
> bmo_deform_utils.cpp.
>
> this two files are in source\blender\bmesh\operators\
>
> What are the directories where the files should put bmo_deform_utils.h
> and bmo_deform_utils.cpp?
> What are policies and style guides in Blender to mix C code and C ++?
>
> Example
>
> bmo_deform_laplacian.c
> //
>
> #Include "bmo_deform_utils.h"
> ...
> au = ftest(4 ,5);
> ...
> ///
>
> bmo_deform_utils.h
> ///
> #ifndef __BMO_DEFORM_UTILS_H__
> #define __BMO_DEFORM_UTILS_H__
>
>
> #ifdef __cplusplus
> extern "C" {
> #endif
>
> int ftest(int x, int y);
>
> #ifdef __cplusplus
> };
> #endif
>
> #endif //__BMO_DEFORM_UTILS_H__
> ///
>
> bmo_deform_utils.cpp
> ///
>
> #include "bmo_deform_utils.h"
>
> class ABC{
> public:
> int a,b;
> ABC(int ax, int by ){
> a = ax;
> b = by;
> }
> int fsum(){
> return a + b;
> }
> };
>
> int ftest(int x, int y){
> ABC A(x, y);
> return A.fsum();
> }
> ///
>
> Thanks in advance.
>
> Att
> Alexander Pinzon Fernandez
> ___
> Bf-committers mailing list
> Bf-committers@blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers
___
Bf-committers mailing list
Bf-committers@blender.org
http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] How to mix C and C++ in Blender
Hi all.
I am working on Bmesh operator for perform laplacian editing for sketch
mesh gsoc project.
If the library to solve the linear system is SuperLU and Eigen3 for matrix
operations, then I must learn how to mix C code (SuperLU) and C ++ (Eigen3)
in Blender. (elegant and organized way).
I have c code file in
source\blender\bmesh\operators\bmo_deform_laplacian.c
in bmo_deform_laplacian.c i call several c++ functions (ej. au = ftest(4
,5);) declare in bmo_deform_utils.h and implemented in
bmo_deform_utils.cpp.
this two files are in source\blender\bmesh\operators\
What are the directories where the files should put bmo_deform_utils.h
and bmo_deform_utils.cpp?
What are policies and style guides in Blender to mix C code and C ++?
Example
bmo_deform_laplacian.c
//
#Include "bmo_deform_utils.h"
...
au = ftest(4 ,5);
...
///
bmo_deform_utils.h
///
#ifndef __BMO_DEFORM_UTILS_H__
#define __BMO_DEFORM_UTILS_H__
#ifdef __cplusplus
extern "C" {
#endif
int ftest(int x, int y);
#ifdef __cplusplus
};
#endif
#endif //__BMO_DEFORM_UTILS_H__
///
bmo_deform_utils.cpp
///
#include "bmo_deform_utils.h"
class ABC{
public:
int a,b;
ABC(int ax, int by ){
a = ax;
b = by;
}
int fsum(){
return a + b;
}
};
int ftest(int x, int y){
ABC A(x, y);
return A.fsum();
}
///
Thanks in advance.
Att
Alexander Pinzon Fernandez
___
Bf-committers mailing list
Bf-committers@blender.org
http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] Windows buildbot naming
Hi, a few people and I just had a discussion in IRC about buildbot being confusing. Imho the current state is a bit problematic, most people will just search for "Windows 64" and download a build. Could we please communicate better that the vc2008 is the official one still and vc2012 and multi view are experimental? Maybe add a label like for MinGW. Or maybe the web UI allows some grouping of "Official Trunk builds" and "Branches / Experimental"? Thanks, Thomas -- Thomas Dinges Blender Developer, Artist and Musician www.dingto.org ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57558] trunk/blender/build_files/scons/ config/win32-vc-config.py: VS 2012 x86 SCons update.
Hi Jürgen, ok this sounds ok, thanks for the info. :) Am 18.06.2013 22:41, schrieb Jürgen Herrmann: > Hi Thomas, > > sorry for beign careless. I changed the compiler flags to reflect the > compiler flags we use in CMake for VS 2008+2012. > > - Debug Builds with SCons were made without /Od and /Ob0 flags > - Release builds were compiled with /O2 /Ob1 , Changed to /O2 /Ob2 as in > CMake > - The usual libs if blocks for VS2012 (not changing vs2008 libs) > > I could enclose the build flags in if-else blocks too but this actually > shouldn't harm VS2008 builds, they work well on my systems. > > -Ursprüngliche Nachricht- > Von: bf-committers-boun...@blender.org > [mailto:bf-committers-boun...@blender.org] Im Auftrag von Thomas Dinges > Gesendet: Dienstag, 18. Juni 2013 22:31 > An: bf-committers@blender.org > Betreff: Re: [Bf-committers] [Bf-blender-cvs] SVN commit: > /data/svn/bf-blender [57558] trunk/blender/build_files/scons/ > config/win32-vc-config.py: VS 2012 x86 SCons update. > > Hi Jürgen, > could you be a bit more precise in your commit messages please? > This does not only change vc2012 related stuff, you also change some > compiler flags which also affect vc2008. > > Thomas > > Am 18.06.2013 22:12, schrieb Juergen Herrmann: >> Revision: 57558 >> > http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision > =57558 >> Author: shadowrom >> Date: 2013-06-18 20:11:58 + (Tue, 18 Jun 2013) >> Log Message: >> --- >> VS 2012 x86 SCons update. >> > -- > Thomas Dinges > Blender Developer, Artist and Musician > > www.dingto.org > > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers > > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers -- Thomas Dinges Blender Developer, Artist and Musician www.dingto.org ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] lattice spline interpolation
David, The Wiki section on submitting patches might be of assistance: http://wiki.blender.org/index.php/Dev:Doc/Process/Patches#Submitting_Patches Nick On Tue, Jun 18, 2013 at 4:27 PM, wrote: > Hello again. I am hoping to get a response this time now that things seem > to have quieted down some on the list. I've got a patch that modifies the > lattice deformer ever so slightly...it simply adds a new spline option. > More info is below. this is a very minor change and I hope it can be > accepted and integrated quickly. We are using Blender at NASA Ames as an > engineering tool, something many of you may never have expected. However, > the lattice deformer is great for doing shape optimization on triangulated > surfaces of airplanes and other aerospace geometries. currently I am using > Blender to do aerostructural analysis, where we predict the deflection of > flexible wings under aerodynamic load (see 787) and actually apply the > deflection via Blender. However, the splines available in the lattice > deformer are inadequate as coded for our use, so I added a new spline...the > Catmull-Rom spline..which is the same as what Blender uses for animation. > this does not remove or alter an > y previous features, it simply adds a another option. > > To whom should I speak about patch files? Thanks. > > > _ > David > > > > > On May 1, 2013, at 3:12 PM, dl...@me.com wrote: > > > Hi all. I discussed updating the lattice spline a few months ago. I > have since done some more research and would like to propose a very simple > improvement to the lattice deformer. The current options are "Linear", > "Cardinal", and "Bspline". The Cardinal spline is set to a tension of 0.71 > which can produce very wavy results when the lattice is dense. However, if > we changed the tension to 0.5, the spline becomes the popular Catmull-Rom > spline and produces very nice results, results we can use here at NASA Ames > to deform wings. The Catmull-Rom spline is already being used in other > parts of Blender, so it's natural to include it as an option in the lattice. > > > > The changes to the code are extremely minor...3-4 source files. 1-2 > lines in most cases, and only 12 or so more in one file with code that > looks very similar to what is already there. I can put together the patch > in about an hour. > > > > My proposal is this: > > > > Add "Catmull-Rom" as an option for the spline interpolators in the > lattice deformer. > > > > > > That's it. Quick and simple. Doesn't alter any current functionality, > it simply expands the current number of options from 3 to 4. > > > > Can such a project be approved? Thanks. > > > > > > > > _ > > David > > > > > > > > > > On Feb 12, 2013, at 3:11 AM, Ton Roosendaal wrote: > > > >> Hi, > >> > >> This code and the formula is like 18 years old, i'd welcome someone > checking on improvements. Rather to make it really good, not just with a > button to set the interplation value. > >> > >> Your 'plot' didnt get through... > >> > >> -Ton- > >> > >> > >> Ton Roosendaal Blender Foundation t...@blender.orgwww.blender.org > >> Blender Institute Entrepotdok 57A 1018AD Amsterdam The Netherlands > >> > >> On 11 Feb, 2013, at 23:26, dl...@mac.com wrote: > >> > >>> Pardon me if this is the wrong list to send this to. If so, I would > appreciate someone pointing me in the right direction. Thanks! > >>> > >>> > >>> The Cardinal spline interpolation for lattices befuddles me. The > tension-like variable (fc in the code) is set to 0.71, which I'm guessing > is approximately 1/sqrt(2). However, this gives very ugly and wavy results > on long straight objects like an airplane wing. In fact, the slopes at the > knots is just plain wrong. However, when fc = 0.5, I get very nice > cubic-spline-like behavior except at the end points where the code logic > just seems to be wrong. Nevertheless, the waviness at the knots is > disconcerting in the current distribution of Blender. The plots below show > what I mean. > >>> > >>> > >>> > >>> > >>> Thoughts? Any chance that fc (tension?) variable can be a parameter > the user could control? > >>> > >>> > >>> _ > >>> David > >>> > >>> > >>> > >>> > >>> ___ > >>> Bf-committers mailing list > >>> Bf-committers@blender.org > >>> http://lists.blender.org/mailman/listinfo/bf-committers > >> > >> ___ > >> Bf-committers mailing list > >> Bf-committers@blender.org > >> http://lists.blender.org/mailman/listinfo/bf-committers > > > > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers > ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57558] trunk/blender/build_files/scons/ config/win32-vc-config.py: VS 2012 x86 SCons update.
Hi Thomas, sorry for beign careless. I changed the compiler flags to reflect the compiler flags we use in CMake for VS 2008+2012. - Debug Builds with SCons were made without /Od and /Ob0 flags - Release builds were compiled with /O2 /Ob1 , Changed to /O2 /Ob2 as in CMake - The usual libs if blocks for VS2012 (not changing vs2008 libs) I could enclose the build flags in if-else blocks too but this actually shouldn't harm VS2008 builds, they work well on my systems. -Ursprüngliche Nachricht- Von: bf-committers-boun...@blender.org [mailto:bf-committers-boun...@blender.org] Im Auftrag von Thomas Dinges Gesendet: Dienstag, 18. Juni 2013 22:31 An: bf-committers@blender.org Betreff: Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57558] trunk/blender/build_files/scons/ config/win32-vc-config.py: VS 2012 x86 SCons update. Hi Jürgen, could you be a bit more precise in your commit messages please? This does not only change vc2012 related stuff, you also change some compiler flags which also affect vc2008. Thomas Am 18.06.2013 22:12, schrieb Juergen Herrmann: > Revision: 57558 > http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision =57558 > Author: shadowrom > Date: 2013-06-18 20:11:58 + (Tue, 18 Jun 2013) > Log Message: > --- > VS 2012 x86 SCons update. > -- Thomas Dinges Blender Developer, Artist and Musician www.dingto.org ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57558] trunk/blender/build_files/scons/ config/win32-vc-config.py: VS 2012 x86 SCons update.
Hi Jürgen, could you be a bit more precise in your commit messages please? This does not only change vc2012 related stuff, you also change some compiler flags which also affect vc2008. Thomas Am 18.06.2013 22:12, schrieb Juergen Herrmann: > Revision: 57558 > > http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57558 > Author: shadowrom > Date: 2013-06-18 20:11:58 + (Tue, 18 Jun 2013) > Log Message: > --- > VS 2012 x86 SCons update. > -- Thomas Dinges Blender Developer, Artist and Musician www.dingto.org ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] lattice spline interpolation
Hello again. I am hoping to get a response this time now that things seem to have quieted down some on the list. I've got a patch that modifies the lattice deformer ever so slightly...it simply adds a new spline option. More info is below. this is a very minor change and I hope it can be accepted and integrated quickly. We are using Blender at NASA Ames as an engineering tool, something many of you may never have expected. However, the lattice deformer is great for doing shape optimization on triangulated surfaces of airplanes and other aerospace geometries. currently I am using Blender to do aerostructural analysis, where we predict the deflection of flexible wings under aerodynamic load (see 787) and actually apply the deflection via Blender. However, the splines available in the lattice deformer are inadequate as coded for our use, so I added a new spline...the Catmull-Rom spline..which is the same as what Blender uses for animation. this does not remove or alter an y previous features, it simply adds a another option. To whom should I speak about patch files? Thanks. _ David On May 1, 2013, at 3:12 PM, dl...@me.com wrote: > Hi all. I discussed updating the lattice spline a few months ago. I have > since done some more research and would like to propose a very simple > improvement to the lattice deformer. The current options are "Linear", > "Cardinal", and "Bspline". The Cardinal spline is set to a tension of 0.71 > which can produce very wavy results when the lattice is dense. However, if > we changed the tension to 0.5, the spline becomes the popular Catmull-Rom > spline and produces very nice results, results we can use here at NASA Ames > to deform wings. The Catmull-Rom spline is already being used in other parts > of Blender, so it's natural to include it as an option in the lattice. > > The changes to the code are extremely minor...3-4 source files. 1-2 lines in > most cases, and only 12 or so more in one file with code that looks very > similar to what is already there. I can put together the patch in about an > hour. > > My proposal is this: > > Add "Catmull-Rom" as an option for the spline interpolators in the lattice > deformer. > > > That's it. Quick and simple. Doesn't alter any current functionality, it > simply expands the current number of options from 3 to 4. > > Can such a project be approved? Thanks. > > > > _ > David > > > > > On Feb 12, 2013, at 3:11 AM, Ton Roosendaal wrote: > >> Hi, >> >> This code and the formula is like 18 years old, i'd welcome someone checking >> on improvements. Rather to make it really good, not just with a button to >> set the interplation value. >> >> Your 'plot' didnt get through... >> >> -Ton- >> >> >> Ton Roosendaal Blender Foundation t...@blender.orgwww.blender.org >> Blender Institute Entrepotdok 57A 1018AD Amsterdam The Netherlands >> >> On 11 Feb, 2013, at 23:26, dl...@mac.com wrote: >> >>> Pardon me if this is the wrong list to send this to. If so, I would >>> appreciate someone pointing me in the right direction. Thanks! >>> >>> >>> The Cardinal spline interpolation for lattices befuddles me. The >>> tension-like variable (fc in the code) is set to 0.71, which I'm guessing >>> is approximately 1/sqrt(2). However, this gives very ugly and wavy results >>> on long straight objects like an airplane wing. In fact, the slopes at the >>> knots is just plain wrong. However, when fc = 0.5, I get very nice >>> cubic-spline-like behavior except at the end points where the code logic >>> just seems to be wrong. Nevertheless, the waviness at the knots is >>> disconcerting in the current distribution of Blender. The plots below show >>> what I mean. >>> >>> >>> >>> >>> Thoughts? Any chance that fc (tension?) variable can be a parameter the >>> user could control? >>> >>> >>> _ >>> David >>> >>> >>> >>> >>> ___ >>> Bf-committers mailing list >>> Bf-committers@blender.org >>> http://lists.blender.org/mailman/listinfo/bf-committers >> >> ___ >> Bf-committers mailing list >> Bf-committers@blender.org >> http://lists.blender.org/mailman/listinfo/bf-committers > ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] VC11 (2012 express for Windows Desktop)
Hi Angus, are you building with CMake or SCons? Building in Debug mode is not supported right now, I am still working on the debug libs and CMake/Scons adaption for VS 2012. But you should be able to build a Release Build with CMake or by calling python.exe scons/scons.py MSVS_VERSION=11.0 /Jürgen -Ursprüngliche Nachricht- Von: bf-committers-boun...@blender.org [mailto:bf-committers-boun...@blender.org] Im Auftrag von Angus Hollands Gesendet: Dienstag, 18. Juni 2013 21:40 An: bf-committers@blender.org Betreff: [Bf-committers] VC11 (2012 express for Windows Desktop) Hi all, I was having trouble compiling for Windows 32 Bit using 2012; It seems that there is a flag error for the compiler in the cycles kernel; "/Zi" "/Ox" (something like that) and also the Python33_d.gz isn't included in the WindowsVC11 libs. Any ideas? ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] VC11 (2012 express for Windows Desktop)
Hi all, I was having trouble compiling for Windows 32 Bit using 2012; It seems that there is a flag error for the compiler in the cycles kernel; "/Zi" "/Ox" (something like that) and also the Python33_d.gz isn't included in the WindowsVC11 libs. Any ideas? ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57543] trunk/blender/intern/cycles: Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree.
I updated my Benchmark with the new builds and the results are astonishing! http://download.shadowrom.de/Benchmark.pdf More than 35% performance gain *yay* -Ursprüngliche Nachricht- Von: bf-committers-boun...@blender.org [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen Herrmann Gesendet: Dienstag, 18. Juni 2013 15:34 An: bf-blender developers Betreff: Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57543] trunk/blender/intern/cycles: Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree. I just tested r57548 (MSVC 2012) on my Core i5 system and got a 23,38% speed up ;) BMW test scene rendered in 530,60 seconds that is nearly the speed of MinGW build r57169 (512,67 seconds) Great work, thanks Brecht! I'll test on my other systems later today. Am 18.06.2013 um 15:26 schrieb Jürgen Herrmann : > Hi Brecht, > > It compiled fine on VS 2012 could you please change your compile fix patch to > compile with the previous settings in MSVC11? > The exception is only needed for VS2008. > > Am 18.06.2013 um 15:21 schrieb Brecht Van Lommel : > >> Looks like a compiler bug. I didn't get that here but I committed >> some more changes to try to fix it. If it doesn't fix things can you >> tell me if this is 32 or 64 bit and cmake or scons? >> >> On Tue, Jun 18, 2013 at 12:09 PM, Thomas Dinges wrote: >>> Hi Brecht, >>> great work, thank you! On Linux I get a nice speedup. :) >>> >>> Windows fails to compile for me though (vc2008). >>> >>> Linking library ==> 'bf_intern_cycles_sse2.lib' >>> kernel_sse3.cpp >>> d:\blender_dev\code\trunk\intern\cycles\kernel\svm\svm_noise.h(38) : >>> fatal error >>> C1001: Internal compile error. >>> (Compilerdatei >>> "f:\dd\vctools\compiler\utc\src\p2\main.c[0x604E:0x00 >>> 00604E]", Zeile 182) >>> >>> Any idea? I tried a full clean build already. >>> >>> Thanks, >>> Thomas >>> >>> Am 18.06.2013 11:36, schrieb Brecht Van Lommel: Revision: 57543 http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57543 Author: blendix Date: 2013-06-18 09:36:06 + (Tue, 18 Jun 2013) Log Message: --- Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree. On the BMW scene, this gives roughly a 10% speedup overall with clang/gcc, and 30% speedup with visual studio (2008). It turns out visual studio was optimizing the existing code quite poorly compared to pretty good autovectorization by clang/gcc, but hand written SSE code also gives a smaller speed boost there. This code isn't enabled when using the hair minimum width feature yet, need to make that work with the SSE code still. >>> >>> -- >>> Thomas Dinges >>> Blender Developer, Artist and Musician >>> >>> www.dingto.org >>> >>> ___ >>> Bf-committers mailing list >>> Bf-committers@blender.org >>> http://lists.blender.org/mailman/listinfo/bf-committers >> ___ >> Bf-committers mailing list >> Bf-committers@blender.org >> http://lists.blender.org/mailman/listinfo/bf-committers > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57555] trunk/blender/source/blender/ blenlib/BLI_path_util.h: Compile fix for r57554 missing include "BLI_string .h" in BLI_path_
Thank you Bastien for the hint.
Fixed in r57556 ;)
-Ursprüngliche Nachricht-
Von: bf-committers-boun...@blender.org
[mailto:bf-committers-boun...@blender.org] Im Auftrag von Bastien Montagne
Gesendet: Dienstag, 18. Juni 2013 21:01
An: bf-committers@blender.org
Betreff: Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender
[57555] trunk/blender/source/blender/ blenlib/BLI_path_util.h: Compile fix for
r57554 missing include "BLI_string .h" in BLI_path_util.h
Hi Juergen,
Not sure this fix is right... We usually avoid including into other includes,
afaik, except for a few specific cases.
As I’m on Linux I can't check it, but right way to go would rather be to add
include of BLI_string.h into the .c files that are newly including
BLI_path_util.h, imho.
At the very least, if we want to make an exception here (granted, it's rather
easy to skip that, esp. as it's only needed for Windows), the BLI_string.h
include should be placed inside an #ifdef WIN32 (and perhaps also moved inside
the extern C block?)...
Cheers,
Bastien
On 18/06/2013 20:42, Juergen Herrmann wrote:
> Revision: 57555
>
> http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57555
> Author: shadowrom
> Date: 2013-06-18 18:42:29 + (Tue, 18 Jun 2013)
> Log Message:
> ---
> Compile fix for r57554 missing include "BLI_string.h" in
> BLI_path_util.h
>
> Revision Links:
> --
>
> http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&re
> vision=57554
>
> Modified Paths:
> --
> trunk/blender/source/blender/blenlib/BLI_path_util.h
>
> Modified: trunk/blender/source/blender/blenlib/BLI_path_util.h
> ===
> --- trunk/blender/source/blender/blenlib/BLI_path_util.h 2013-06-18
> 18:11:52 UTC (rev 57554)
> +++ trunk/blender/source/blender/blenlib/BLI_path_util.h 2013-06-18
> 18:42:29 UTC (rev 57555)
> @@ -31,6 +31,8 @@
>* \ingroup bli
>*/
>
> +#include "BLI_string.h"
> +
> #ifdef __cplusplus
> extern "C" {
> #endif
> @@ -162,7 +164,7 @@
> bool BLI_path_is_rel(const char *path);
>
> /* path string comparisons: case-insensitive for Windows,
> case-sensitive otherwise */ -#ifdef WIN32
> +#if defined(WIN32)
> # define BLI_path_cmp BLI_strcasecmp
> # define BLI_path_ncmp BLI_strncasecmp
> #else
>
> ___
> Bf-blender-cvs mailing list
> bf-blender-...@blender.org
> http://lists.blender.org/mailman/listinfo/bf-blender-cvs
>
___
Bf-committers mailing list
Bf-committers@blender.org
http://lists.blender.org/mailman/listinfo/bf-committers
___
Bf-committers mailing list
Bf-committers@blender.org
http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57555] trunk/blender/source/blender/ blenlib/BLI_path_util.h: Compile fix for r57554 missing include "BLI_string .h" in BLI_path_
Hi Juergen,
Not sure this fix is right... We usually avoid including into other
includes, afaik, except for a few specific cases.
As I’m on Linux I can't check it, but right way to go would rather be to
add include of BLI_string.h into the .c files that are newly including
BLI_path_util.h, imho.
At the very least, if we want to make an exception here (granted, it's
rather easy to skip that, esp. as it's only needed for Windows), the
BLI_string.h include should be placed inside an #ifdef WIN32 (and
perhaps also moved inside the extern C block?)...
Cheers,
Bastien
On 18/06/2013 20:42, Juergen Herrmann wrote:
> Revision: 57555
>
> http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57555
> Author: shadowrom
> Date: 2013-06-18 18:42:29 + (Tue, 18 Jun 2013)
> Log Message:
> ---
> Compile fix for r57554 missing include "BLI_string.h" in BLI_path_util.h
>
> Revision Links:
> --
>
> http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57554
>
> Modified Paths:
> --
> trunk/blender/source/blender/blenlib/BLI_path_util.h
>
> Modified: trunk/blender/source/blender/blenlib/BLI_path_util.h
> ===
> --- trunk/blender/source/blender/blenlib/BLI_path_util.h 2013-06-18
> 18:11:52 UTC (rev 57554)
> +++ trunk/blender/source/blender/blenlib/BLI_path_util.h 2013-06-18
> 18:42:29 UTC (rev 57555)
> @@ -31,6 +31,8 @@
>* \ingroup bli
>*/
>
> +#include "BLI_string.h"
> +
> #ifdef __cplusplus
> extern "C" {
> #endif
> @@ -162,7 +164,7 @@
> bool BLI_path_is_rel(const char *path);
>
> /* path string comparisons: case-insensitive for Windows, case-sensitive
> otherwise */
> -#ifdef WIN32
> +#if defined(WIN32)
> # define BLI_path_cmp BLI_strcasecmp
> # define BLI_path_ncmp BLI_strncasecmp
> #else
>
> ___
> Bf-blender-cvs mailing list
> bf-blender-...@blender.org
> http://lists.blender.org/mailman/listinfo/bf-blender-cvs
>
___
Bf-committers mailing list
Bf-committers@blender.org
http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Wayland and Blender
2013/6/18 Sergej Reich : > Am Dienstag, den 18.06.2013, 11:03 -0300 schrieb Wander Lairson Costa: >> 2013/6/18 Campbell Barton : >> > >> > Quite sure nobody started this yet, I was interested to do this but >> > couldn't manage to get any OpenGL apps running on weston (even >> > weston-glxgears). >> > >> >> That one of issues I need to ask on wayland-dev. This is due Weston >> using OpenGL ES, right? Full OpenGL on Linux depends on glx, which >> depends on X11 libraries. If this problem isn't fixed yet, Weston now >> can run under X11, so maybe I can test my implementation (when I have >> something to test) in this way, am I right? > > I'm pretty sure EGL which wayland uses instead of GLX works for regular > Gl as well. So we'd have to use that instead. > Yes, I am using EGL for windowing setup. [snip] -- Best Regards, Wander Lairson Costa ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Wayland and Blender
Am Dienstag, den 18.06.2013, 11:03 -0300 schrieb Wander Lairson Costa: > 2013/6/18 Campbell Barton : > > > > Quite sure nobody started this yet, I was interested to do this but > > couldn't manage to get any OpenGL apps running on weston (even > > weston-glxgears). > > > > That one of issues I need to ask on wayland-dev. This is due Weston > using OpenGL ES, right? Full OpenGL on Linux depends on glx, which > depends on X11 libraries. If this problem isn't fixed yet, Weston now > can run under X11, so maybe I can test my implementation (when I have > something to test) in this way, am I right? I'm pretty sure EGL which wayland uses instead of GLX works for regular Gl as well. So we'd have to use that instead. > > > Would be great if you could post on this list if you have a repo with > > some basics working. > > I just have started, there is almost nothing relevant yet. As soon as > I have something to show, I will make my repo public and post the link > here. > > ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Wayland and Blender
2013/6/18 Campbell Barton : > > Quite sure nobody started this yet, I was interested to do this but > couldn't manage to get any OpenGL apps running on weston (even > weston-glxgears). > That one of issues I need to ask on wayland-dev. This is due Weston using OpenGL ES, right? Full OpenGL on Linux depends on glx, which depends on X11 libraries. If this problem isn't fixed yet, Weston now can run under X11, so maybe I can test my implementation (when I have something to test) in this way, am I right? > Would be great if you could post on this list if you have a repo with > some basics working. I just have started, there is almost nothing relevant yet. As soon as I have something to show, I will make my repo public and post the link here. -- Best Regards, Wander Lairson Costa ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Wayland and Blender
On Tue, Jun 18, 2013 at 10:08 PM, Wander Lairson Costa wrote: > Hi, > > I would like to know if there are any efforts for Wayland going on. I > did a search on web and saw some past discussions on wayland-devel > list. > > I just started to add Wayland support this weekend, but the work is > rather initial and will take sometime until I have something worthy to > go public, as I am doing it in my spare time and I am still learning > the ins and outs of Wayland. > > -- > Best Regards, > Wander Lairson Costa Quite sure nobody started this yet, I was interested to do this but couldn't manage to get any OpenGL apps running on weston (even weston-glxgears). Would be great if you could post on this list if you have a repo with some basics working. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57543] trunk/blender/intern/cycles: Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree.
I just tested r57548 (MSVC 2012) on my Core i5 system and got a 23,38% speed up ;) BMW test scene rendered in 530,60 seconds that is nearly the speed of MinGW build r57169 (512,67 seconds) Great work, thanks Brecht! I'll test on my other systems later today. Am 18.06.2013 um 15:26 schrieb Jürgen Herrmann : > Hi Brecht, > > It compiled fine on VS 2012 could you please change your compile fix patch to > compile with the previous settings in MSVC11? > The exception is only needed for VS2008. > > Am 18.06.2013 um 15:21 schrieb Brecht Van Lommel : > >> Looks like a compiler bug. I didn't get that here but I committed some >> more changes to try to fix it. If it doesn't fix things can you tell >> me if this is 32 or 64 bit and cmake or scons? >> >> On Tue, Jun 18, 2013 at 12:09 PM, Thomas Dinges wrote: >>> Hi Brecht, >>> great work, thank you! On Linux I get a nice speedup. :) >>> >>> Windows fails to compile for me though (vc2008). >>> >>> Linking library ==> 'bf_intern_cycles_sse2.lib' >>> kernel_sse3.cpp >>> d:\blender_dev\code\trunk\intern\cycles\kernel\svm\svm_noise.h(38) : >>> fatal error >>> C1001: Internal compile error. >>> (Compilerdatei >>> "f:\dd\vctools\compiler\utc\src\p2\main.c[0x604E:0x00 >>> 00604E]", Zeile 182) >>> >>> Any idea? I tried a full clean build already. >>> >>> Thanks, >>> Thomas >>> >>> Am 18.06.2013 11:36, schrieb Brecht Van Lommel: Revision: 57543 http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57543 Author: blendix Date: 2013-06-18 09:36:06 + (Tue, 18 Jun 2013) Log Message: --- Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree. On the BMW scene, this gives roughly a 10% speedup overall with clang/gcc, and 30% speedup with visual studio (2008). It turns out visual studio was optimizing the existing code quite poorly compared to pretty good autovectorization by clang/gcc, but hand written SSE code also gives a smaller speed boost there. This code isn't enabled when using the hair minimum width feature yet, need to make that work with the SSE code still. >>> >>> -- >>> Thomas Dinges >>> Blender Developer, Artist and Musician >>> >>> www.dingto.org >>> >>> ___ >>> Bf-committers mailing list >>> Bf-committers@blender.org >>> http://lists.blender.org/mailman/listinfo/bf-committers >> ___ >> Bf-committers mailing list >> Bf-committers@blender.org >> http://lists.blender.org/mailman/listinfo/bf-committers > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57543] trunk/blender/intern/cycles: Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree.
Hi Brecht, It compiled fine on VS 2012 could you please change your compile fix patch to compile with the previous settings in MSVC11? The exception is only needed for VS2008. Am 18.06.2013 um 15:21 schrieb Brecht Van Lommel : > Looks like a compiler bug. I didn't get that here but I committed some > more changes to try to fix it. If it doesn't fix things can you tell > me if this is 32 or 64 bit and cmake or scons? > > On Tue, Jun 18, 2013 at 12:09 PM, Thomas Dinges wrote: >> Hi Brecht, >> great work, thank you! On Linux I get a nice speedup. :) >> >> Windows fails to compile for me though (vc2008). >> >> Linking library ==> 'bf_intern_cycles_sse2.lib' >> kernel_sse3.cpp >> d:\blender_dev\code\trunk\intern\cycles\kernel\svm\svm_noise.h(38) : >> fatal error >> C1001: Internal compile error. >> (Compilerdatei >> "f:\dd\vctools\compiler\utc\src\p2\main.c[0x604E:0x00 >> 00604E]", Zeile 182) >> >> Any idea? I tried a full clean build already. >> >> Thanks, >> Thomas >> >> Am 18.06.2013 11:36, schrieb Brecht Van Lommel: >>> Revision: 57543 >>> >>> http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57543 >>> Author: blendix >>> Date: 2013-06-18 09:36:06 + (Tue, 18 Jun 2013) >>> Log Message: >>> --- >>> Cycles: optimization for BVH traveral on CPU's with SSE3, using code from >>> Embree. >>> >>> On the BMW scene, this gives roughly a 10% speedup overall with clang/gcc, >>> and 30% >>> speedup with visual studio (2008). It turns out visual studio was >>> optimizing the >>> existing code quite poorly compared to pretty good autovectorization by >>> clang/gcc, >>> but hand written SSE code also gives a smaller speed boost there. >>> >>> This code isn't enabled when using the hair minimum width feature yet, need >>> to >>> make that work with the SSE code still. >> >> -- >> Thomas Dinges >> Blender Developer, Artist and Musician >> >> www.dingto.org >> >> ___ >> Bf-committers mailing list >> Bf-committers@blender.org >> http://lists.blender.org/mailman/listinfo/bf-committers > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57543] trunk/blender/intern/cycles: Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree.
Looks like a compiler bug. I didn't get that here but I committed some more changes to try to fix it. If it doesn't fix things can you tell me if this is 32 or 64 bit and cmake or scons? On Tue, Jun 18, 2013 at 12:09 PM, Thomas Dinges wrote: > Hi Brecht, > great work, thank you! On Linux I get a nice speedup. :) > > Windows fails to compile for me though (vc2008). > > Linking library ==> 'bf_intern_cycles_sse2.lib' > kernel_sse3.cpp > d:\blender_dev\code\trunk\intern\cycles\kernel\svm\svm_noise.h(38) : > fatal error > C1001: Internal compile error. > (Compilerdatei > "f:\dd\vctools\compiler\utc\src\p2\main.c[0x604E:0x00 > 00604E]", Zeile 182) > > Any idea? I tried a full clean build already. > > Thanks, > Thomas > > Am 18.06.2013 11:36, schrieb Brecht Van Lommel: >> Revision: 57543 >> >> http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57543 >> Author: blendix >> Date: 2013-06-18 09:36:06 + (Tue, 18 Jun 2013) >> Log Message: >> --- >> Cycles: optimization for BVH traveral on CPU's with SSE3, using code from >> Embree. >> >> On the BMW scene, this gives roughly a 10% speedup overall with clang/gcc, >> and 30% >> speedup with visual studio (2008). It turns out visual studio was optimizing >> the >> existing code quite poorly compared to pretty good autovectorization by >> clang/gcc, >> but hand written SSE code also gives a smaller speed boost there. >> >> This code isn't enabled when using the hair minimum width feature yet, need >> to >> make that work with the SSE code still. > > -- > Thomas Dinges > Blender Developer, Artist and Musician > > www.dingto.org > > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Cycles CPU Kernel on MSVC *all versions*
Hi Brecht, Cool stuff ;) I submitted a little Build patch for VS2012 to include intrin.h which is needed for _BitScanForward an _BitScanReverse I am curious what vs2012 makes of this. Am 18.06.2013 um 13:59 schrieb Brecht Van Lommel : > Hi Jürgen , > > BVH / triangle intersection is always going to take up a major part of > rendering time, but I managed to optimize it some. For MSVC the BVH > intersection is now twice as fast for e.g. the BMW scene, which gives > a 30% render time reduction overall. I also added the /Ox compiler > flag (which contains the other /O compiler flags you posted for the > benchmark), that helps maybe 2% here. > > On Fri, Jun 7, 2013 at 9:32 AM, Jürgen Herrmann wrote: >> Hi there, >> >> after the Benchmarking I did I decided to try analyze the diferences between >> MinGW and MSVC build in depth. >> First of all I ran several Benchmarks of math functions in both GCC an MSVC >> and found out: >> Amazing! they are both equally fast! So math functions isn't the bottleneck >> here. >> So I tried to get a grip on real data and tried out "Very Sleepy" ( free >> download at: http://www.codersnotes.com/sleepy ) to profile cycles on MSVC. >> >> What I found here was quite interresting: >> While rendering the most time is spent in ccl::bvh_intersect_instancing with >> 1292.15s (53.70%) O_o >> Within this call the most time is spent in ccl::bvh_node_intersect with >> 872.47s (36.26%) . >> >> So this is the part we'll have to look at ... I wonder how MinGW builds >> perform here, but I don't have a proper optimized build with debug info to >> compare :/ >> >> /Jürgen >> ___ >> Bf-committers mailing list >> Bf-committers@blender.org >> http://lists.blender.org/mailman/listinfo/bf-committers > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] Wayland and Blender
Hi, I would like to know if there are any efforts for Wayland going on. I did a search on web and saw some past discussions on wayland-devel list. I just started to add Wayland support this weekend, but the work is rather initial and will take sometime until I have something worthy to go public, as I am doing it in my spare time and I am still learning the ins and outs of Wayland. -- Best Regards, Wander Lairson Costa ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Cycles CPU Kernel on MSVC *all versions*
Hi Jürgen , BVH / triangle intersection is always going to take up a major part of rendering time, but I managed to optimize it some. For MSVC the BVH intersection is now twice as fast for e.g. the BMW scene, which gives a 30% render time reduction overall. I also added the /Ox compiler flag (which contains the other /O compiler flags you posted for the benchmark), that helps maybe 2% here. On Fri, Jun 7, 2013 at 9:32 AM, Jürgen Herrmann wrote: > Hi there, > > after the Benchmarking I did I decided to try analyze the diferences between > MinGW and MSVC build in depth. > First of all I ran several Benchmarks of math functions in both GCC an MSVC > and found out: > Amazing! they are both equally fast! So math functions isn't the bottleneck > here. > So I tried to get a grip on real data and tried out "Very Sleepy" ( free > download at: http://www.codersnotes.com/sleepy ) to profile cycles on MSVC. > > What I found here was quite interresting: > While rendering the most time is spent in ccl::bvh_intersect_instancing with > 1292.15s (53.70%) O_o > Within this call the most time is spent in ccl::bvh_node_intersect with > 872.47s (36.26%) . > > So this is the part we'll have to look at ... I wonder how MinGW builds > perform here, but I don't have a proper optimized build with debug info to > compare :/ > > /Jürgen > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
Hi Brecht,
I'll change that and commit the change ASAP.
Thanks for review ;)
Am 18.06.2013 um 11:50 schrieb Brecht Van Lommel :
> Looks good to me, please commit!
>
> Maybe make one change, replacing
> if builder.endswith('win64_scons_vc2012') or
> builder.endswith('win32_scons_vc2012'):
> by:
> if builder.endswith('vc2012'):
>
> On Mon, Jun 17, 2013 at 10:06 PM, Jürgen Herrmann wrote:
>> Ok,
>>
>> i am tired, postet the wrong version oft he patch :(
>>
>> The porper versin is here:
>>
>> http://www.pasteall.org/43250
>>
>> I'll go to bed now ^_^
>> The build bot is tested locally and can be started as soon as scons is
>> patched ;)
>>
>> -Ursprüngliche Nachricht-
>> Von: bf-committers-boun...@blender.org
>> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen Herrmann
>> Gesendet: Montag, 17. Juni 2013 21:46
>> An: 'bf-blender developers'
>> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>>
>> Alright, this little problem bugged me, so I fixed it!
>>
>> I got rid of the second config file and found a way to detect the used MSVC
>> Version within the config.
>>
>> Patch for scons and buildbot attached ;)
>>
>> http://www.pasteall.org/43249
>>
>>
>>
>> -Ursprüngliche Nachricht-
>> Von: bf-committers-boun...@blender.org
>> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen Herrmann
>> Gesendet: Montag, 17. Juni 2013 20:14
>> An: 'bf-blender developers'
>> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>>
>> Hi again,
>>
>> i prepared buildbot and scons fort he buildbot builds.
>> Now buildbot will create a package named
>>
>> blender-VERSION-REVISION-PLATFORM-vc11.zip
>>
>> for VC 2012 builds.
>> Compiling and Packaging works well now, I'll try to start the Buildbot for
>> x64 as soon as the patch is approved and commited ;)
>> X86 will follow after that.
>>
>> /Jürgen
>>
>> -Ursprüngliche Nachricht-
>> Von: bf-committers-boun...@blender.org
>> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen Herrmann
>> Gesendet: Montag, 17. Juni 2013 19:16
>> An: 'bf-blender developers'
>> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>>
>> Hi Brecht,
>>
>> I tried that by usingf
>> if env['MSVC_VERSION'] == '11.0':
>>
>> But it seems that "env" is not present in the scope of this config file, it
>> always fails :(
>>
>> The MacOSX config file uses system commands to determine their versions, but
>> this is quite hard for windows.
>>
>> /Jürgen
>>
>> -Ursprüngliche Nachricht-
>> Von: bf-committers-boun...@blender.org
>> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Brecht Van Lommel
>> Gesendet: Montag, 17. Juni 2013 19:12
>> An: bf-blender developers
>> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>>
>> It looks good to me, except that perhaps we can avoid copying the config
>> file. Can we add if/else in the win64-vc-config.py based on the MSVS
>> version? It's only 4 lines or so that are actually different.
>>
>> Brecht.
>>
>> On Mon, Jun 17, 2013 at 6:43 PM, Jürgen Herrmann wrote:
>>> Sorry, my fault ... getting tired ;)
>>>
>>> http://www.pasteall.org/43245
>>>
>>>
>>>
>>> -Ursprüngliche Nachricht-
>>> Von: bf-committers-boun...@blender.org
>>> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Thomas
>>> Dinges
>>> Gesendet: Montag, 17. Juni 2013 18:39
>>> An: bf-blender developers
>>> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>>>
>>> Hi Jürgen,
>>> the patch didn't get attached, maybe upload it to pasteall.org.
>>>
>>> Thanks,
>>> Thomas
>>>
>>> Am 17.06.2013 18:37, schrieb Jürgen Herrmann:
Hi Brecht and Thomas,
I attached my patch for SCons.
I tested this with VC 2008 and 2012 and MinGW64.
I don't have other systems to test right now :/ It works quite well
though this is a little hacky because I don't get into the internals
of
>>> SCons that good.
I set it to default to MSVC 2008 so
- calling "py scons/scons.py" will build with MSVC2008
- calling "py scons\scons.py MSVS_VERSION=11.0" will build with VS
2012
- calling "py scons/scons.py BF_TOOLSET=mingw" builds MinGW as usual.
Could you two have a look at this please, I bet you'll have some
ideas how
>>> to make it better.
/Jürgen
-Ursprüngliche Nachricht-
Von: bf-committers-boun...@blender.org
[mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen
Herrmann
Gesendet: Montag, 17. Juni 2013 14:50
An: bf-blender developers
Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
Thanks Brecht ;)
I'll try to get this running this evening.
I'll still have to look into scons build.
I got it to work by querying env['MSVS_VERSION'] which is
automatically
>>> set to 11.0 when vs2012 is installed.
But I'll have to test in anothe
[Bf-committers] Request for patch review: quick hack referring to select more/less operators
Hi all. Yesterday I submitted a patch in the patch tracker (#35789) which implements the functionality requested for the first quick hack of this page: http://wiki.blender.org/index.php/Dev:Doc/Quick_Hacks, which is select more/less operators for both armatures and lattices. So, I would like to request a review for that patch. Since I'm new on this, please do not doubt to tell me how could I improve the information given in the tracker, or any other kind of mistake I could have done. Best regards. Pedro. -- When the egg is broken from outside, life ends. When the egg is broken from inside, life begins. Great things always begin *from inside*. es.linkedin.com/in/rieramartorell/ ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] BGE Future
Hi Gavin, While I tend to agree on your observations I do think the direction for the realtime part of Blender (BGE) should be a more general one. So not necessary a game direction. A game engine is just an engine optimised for realtime applications. So if Blender puts more focus into the realtime possibilities it will open the way for way more artists and designers than only the game artists. i.e.: * Data Visualisation * (Scientific) simulations * Generative applications * interactive installations (new media art) * generative animations for use in animations So this 'interactive' focus includes possibilies for games while giving Blender a 'game' focus would potentially exclude other options. This is what we suffer a lot for example. The game engine is focused on game application which makes us do a lot of tricks to get it to do more. By the way modern games are also getting more out of the computers using more Mixed Reality technologies and new HID devices. Take a Kinect for example... A kinect device can do way more than only controlling a game, see puppeteering examples... So I for one am very happy which this focus and expect to see more unification between BGE and Blender. So I'm quite sure the BGE will benefit from it. I tend to agree on Ton's statements in the past i.e.: http://blenderartists.org/forum/showthread.php?226524-Where-is-the-development#10 I just hope the blenderplayer (or similar headless mode) will remain, but I can hardy imagine without it anyway. The future is more with realtime/interactive possibilities which includes games, not vice-versa. Rg, Arnaud -- Stichting z25.org Concordiastraat 67A 3551 EM Utrecht The Netherlands On Mon, 17 Jun 2013 08:30:26 -0600 Gavin Howard wrote: > All, > > I'm starting a new thread because it looks like the previous thread > got mixed in with another thread. > > I never thought that I would write a post like this, but quite > frankly, I was surprised by how quickly user opposition cropped up > from Ton's proposed changes to the BGE. At this point, I feel that I > need to speak frankly. I hope that I don't step on anyone's toes, and > I also hope that I don't offend anyone. I will be writing down the > situation as I understand it (so I will be repeating things everyone > already knows), and I will be offering my commentary and opinion. That > said, here goes. > > I started using Blender regularly about a year and a half ago. From > the beginning, I started regularly observing the BlenderNation forums, > as well as other gathering places for users and devs. Even at the > start, I saw a little bit of tension between BGE users and Blender > devs. The users LOVE their engine, maybe more so than regular Blender > users love Blender. They desperately want some dev time put in the > BGE, and the devs just haven't had time or interest. > > Obviously, a change has been needed. And then, Ton makes proposed > changes that sound as though the BGE is going away as BGE, even if > there is no loss of functionality. Now, I want it understood that I > have never been a BGE user. I don't have any use for it because I > don't make, or even play, video games, but there are quite a few BGE > users that want to keep the engine an engine. So they toss around the > option of creating a fork from an existing build that has many > user-submitted patches applied. > > To complicate things, Daniel Stokes has a BGE project. He is now > working on an engine that may not be an engine in a year or two. > (Sorry, Daniel!) Nevertheless, it was time for Ton to put out the > roadmap. I believe the only mistake he made was that he didn't > anticipate how much the BGE means to its users. But it was definitely > time. Version 2.68 is more than halfway done, and 2.69 is just around > the corner. We need to plan for 2.7x, and we need to do it soon. > > As an up-and-coming animator, being able to apply logic nodes to > animations sounds incredibly good. Having an interactive mode that is > not a game engine sounds incredibly good. Being able to do rule-based > animation (for crowds and things like that) sounds AWESOMELY good. I > LIKE the direction that Ton wants to go. Blender's main purpose is to > produce images and animations, so it fits with the philosophy as well. > However, I also understand that many people see the game engine for > what it is: a game engine. > > So, here's MY proposal, if it even matters. I propose that Daniel keep > his project. Yes, I know, it may not matter in the long run. But wait > a second! His project is to add level-of-detail support to the BGE. As > an animator, if my software can automatically adjust the level of > detail for objects based on distance from the camera, I would be very > happy, so even if BGE disappears, that code won't, which means that > mainstream Blender would get that capability. That is VERY cool. And > then, his project is to do a lot of refactoring, bug fixing, and so > on. Well, the BGE apparently n
Re: [Bf-committers] [Bf-blender-cvs] SVN commit: /data/svn/bf-blender [57543] trunk/blender/intern/cycles: Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree.
Hi Brecht, great work, thank you! On Linux I get a nice speedup. :) Windows fails to compile for me though (vc2008). Linking library ==> 'bf_intern_cycles_sse2.lib' kernel_sse3.cpp d:\blender_dev\code\trunk\intern\cycles\kernel\svm\svm_noise.h(38) : fatal error C1001: Internal compile error. (Compilerdatei "f:\dd\vctools\compiler\utc\src\p2\main.c[0x604E:0x00 00604E]", Zeile 182) Any idea? I tried a full clean build already. Thanks, Thomas Am 18.06.2013 11:36, schrieb Brecht Van Lommel: > Revision: 57543 > > http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=57543 > Author: blendix > Date: 2013-06-18 09:36:06 + (Tue, 18 Jun 2013) > Log Message: > --- > Cycles: optimization for BVH traveral on CPU's with SSE3, using code from > Embree. > > On the BMW scene, this gives roughly a 10% speedup overall with clang/gcc, > and 30% > speedup with visual studio (2008). It turns out visual studio was optimizing > the > existing code quite poorly compared to pretty good autovectorization by > clang/gcc, > but hand written SSE code also gives a smaller speed boost there. > > This code isn't enabled when using the hair minimum width feature yet, need to > make that work with the SSE code still. -- Thomas Dinges Blender Developer, Artist and Musician www.dingto.org ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
Looks good to me, please commit!
Maybe make one change, replacing
if builder.endswith('win64_scons_vc2012') or
builder.endswith('win32_scons_vc2012'):
by:
if builder.endswith('vc2012'):
On Mon, Jun 17, 2013 at 10:06 PM, Jürgen Herrmann wrote:
> Ok,
>
> i am tired, postet the wrong version oft he patch :(
>
> The porper versin is here:
>
> http://www.pasteall.org/43250
>
> I'll go to bed now ^_^
> The build bot is tested locally and can be started as soon as scons is
> patched ;)
>
> -Ursprüngliche Nachricht-
> Von: bf-committers-boun...@blender.org
> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen Herrmann
> Gesendet: Montag, 17. Juni 2013 21:46
> An: 'bf-blender developers'
> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>
> Alright, this little problem bugged me, so I fixed it!
>
> I got rid of the second config file and found a way to detect the used MSVC
> Version within the config.
>
> Patch for scons and buildbot attached ;)
>
> http://www.pasteall.org/43249
>
>
>
> -Ursprüngliche Nachricht-
> Von: bf-committers-boun...@blender.org
> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen Herrmann
> Gesendet: Montag, 17. Juni 2013 20:14
> An: 'bf-blender developers'
> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>
> Hi again,
>
> i prepared buildbot and scons fort he buildbot builds.
> Now buildbot will create a package named
>
> blender-VERSION-REVISION-PLATFORM-vc11.zip
>
> for VC 2012 builds.
> Compiling and Packaging works well now, I'll try to start the Buildbot for
> x64 as soon as the patch is approved and commited ;)
> X86 will follow after that.
>
> /Jürgen
>
> -Ursprüngliche Nachricht-
> Von: bf-committers-boun...@blender.org
> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen Herrmann
> Gesendet: Montag, 17. Juni 2013 19:16
> An: 'bf-blender developers'
> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>
> Hi Brecht,
>
> I tried that by usingf
> if env['MSVC_VERSION'] == '11.0':
>
> But it seems that "env" is not present in the scope of this config file, it
> always fails :(
>
> The MacOSX config file uses system commands to determine their versions, but
> this is quite hard for windows.
>
> /Jürgen
>
> -Ursprüngliche Nachricht-
> Von: bf-committers-boun...@blender.org
> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Brecht Van Lommel
> Gesendet: Montag, 17. Juni 2013 19:12
> An: bf-blender developers
> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>
> It looks good to me, except that perhaps we can avoid copying the config
> file. Can we add if/else in the win64-vc-config.py based on the MSVS
> version? It's only 4 lines or so that are actually different.
>
> Brecht.
>
> On Mon, Jun 17, 2013 at 6:43 PM, Jürgen Herrmann wrote:
>> Sorry, my fault ... getting tired ;)
>>
>> http://www.pasteall.org/43245
>>
>>
>>
>> -Ursprüngliche Nachricht-
>> Von: bf-committers-boun...@blender.org
>> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Thomas
>> Dinges
>> Gesendet: Montag, 17. Juni 2013 18:39
>> An: bf-blender developers
>> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>>
>> Hi Jürgen,
>> the patch didn't get attached, maybe upload it to pasteall.org.
>>
>> Thanks,
>> Thomas
>>
>> Am 17.06.2013 18:37, schrieb Jürgen Herrmann:
>>> Hi Brecht and Thomas,
>>>
>>> I attached my patch for SCons.
>>> I tested this with VC 2008 and 2012 and MinGW64.
>>> I don't have other systems to test right now :/ It works quite well
>>> though this is a little hacky because I don't get into the internals
>>> of
>> SCons that good.
>>> I set it to default to MSVC 2008 so
>>> - calling "py scons/scons.py" will build with MSVC2008
>>> - calling "py scons\scons.py MSVS_VERSION=11.0" will build with VS
>>> 2012
>>> - calling "py scons/scons.py BF_TOOLSET=mingw" builds MinGW as usual.
>>>
>>> Could you two have a look at this please, I bet you'll have some
>>> ideas how
>> to make it better.
>>>
>>> /Jürgen
>>>
>>> -Ursprüngliche Nachricht-
>>> Von: bf-committers-boun...@blender.org
>>> [mailto:bf-committers-boun...@blender.org] Im Auftrag von Jürgen
>>> Herrmann
>>> Gesendet: Montag, 17. Juni 2013 14:50
>>> An: bf-blender developers
>>> Betreff: Re: [Bf-committers] Buildbot setup for MSVC 2012 testbuilds
>>>
>>> Thanks Brecht ;)
>>> I'll try to get this running this evening.
>>> I'll still have to look into scons build.
>>> I got it to work by querying env['MSVS_VERSION'] which is
>>> automatically
>> set to 11.0 when vs2012 is installed.
>>> But I'll have to test in another machine without VS2012 and with
>>> VS2008 if
>> it doesn't break things.
>>> And with MinGW...
>>> After that test I'll commit the scons changes and set up the buildbot
>>> ;)
>>>
>>> /Jürgen
>>>
>>> Am 17.06.2013 um 14:23 schrieb Brecht Van Lommel
>> :
>>>
I've sent you a mail with all the details.
On Sun, Ju
Re: [Bf-committers] Multiview branch feedback
On 18/06/2013 03:07, Matt Ebb wrote: > On Tue, Jun 18, 2013 at 10:55 AM, Dalai Felinto wrote: > >> All the per-stream tweak can be accomplished with the "View Switch" >> node. e.g., to set the convergence you can do: >> >> http://dalaifelinto.com/ftp/tmp/multiview_convergence1.jpg >> http://dalaifelinto.com/ftp/tmp/multiview_convergence2.jpg > > Is it possible to split the individual left and right views out of a > multiview stream into two single view streams though? In that network it > looks like you're applying two translations to two multi-view streams. > ___ > Bf-committers mailing list > Bf-committers@blender.org > http://lists.blender.org/mailman/listinfo/bf-committers Maybe I am wrong but are translations really sufficient to correspond to convergence. I am wondering if a solution with UVs of mesh planes used with Map UV node as compositing planes would not be better. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Multiview branch feedback
On 18/06/2013 00:05, Dalai Felinto wrote: > That said, the planes as a visual feedback to the "depth box" will be > supported in the bultin stereo camera. I'm not sure if they should be > directly manipulated though (as in Bartek's addon). But I will look at > that during the builtin stereo camera development. I cared about visual feedback in 3D view. My expectation is an automatic refreshing of this visual feedback at stereo camera settings change. We have one for Limits and Mist options in Display panel of standard camera. Thank you for the answers Brecht and Dalai. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
