Re: [Bf-committers] Blender and OpenCL

2010-09-02 Thread Nathan Letwory
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 1.9.2010 23:25, Jeroen Bakker wrote: > Hi, > > I have created an OpenCL implementation of the bokeh blur. I got a > speedup (2 times faster) on my old hardware, but also some stability > issues. I think that some issues concern my old NVidia + Op

Re: [Bf-committers] Blender and OpenCL

2010-09-02 Thread Lukas Tönne
experience can be quality blurs(especially >> defocus), UV remap, >> bilateral blur. >> With these the advantages would be visible even with the bus problems. >> With regards, >> Vilem Novak >> >> >>>

Re: [Bf-committers] Blender and OpenCL

2010-09-01 Thread Jeroen Bakker
;> Původní zpráva >> Od: Jeroen Bakker >> Předmět: Re: [Bf-committers] Blender and OpenCL >> Datum: 29.8.2010 11:19:15 >> >> Hi Lukas, >> >> Your explanation is a good one. Didn't come up to wr

Re: [Bf-committers] Blender and OpenCL

2010-08-30 Thread François T .
t; bilateral blur. > With these the advantages would be visible even with the bus problems. > With regards, > Vilem Novak > > > Původní zpráva > > Od: Jeroen Bakker > > Předmět: Re: [Bf-committers] Blende

Re: [Bf-committers] Blender and OpenCL

2010-08-29 Thread Vilem Novak
> Původní zpráva > Od: Jeroen Bakker > Předmět: Re: [Bf-committers] Blender and OpenCL > Datum: 29.8.2010 11:19:15 > > Hi Lukas, > > Your explanation is a good one. Didn't come up to write it down that way. &

Re: [Bf-committers] Blender and OpenCL

2010-08-29 Thread Jeroen Bakker
On 08/29/2010 02:02 AM, Ralph Giles wrote: > 2010/8/28 Arturo José Pérez Verdú: > > >> #if defined(__APPLE__) || defined(__MACOSX) >> #include >> #else >> #include >> #endif >> > Yes, sadly this bit of incompatibility was willfully propagated from OpenGL. > :( > > -r >

Re: [Bf-committers] Blender and OpenCL

2010-08-29 Thread Jeroen Bakker
Hi Lukas, Your explanation is a good one. Didn't come up to write it down that way. The issue with memory during compositing is the way the nodes-editor works. When changing a node-value (like degree) only the rotate-node and all dependent nodes are re-calculated. The input-image is not re-calc

Re: [Bf-committers] Blender and OpenCL

2010-08-28 Thread Alex Fraser
On Sun, Aug 29, 2010 at 5:50 AM, Ralph Giles wrote: > > But as Jeroen points out, OpenCL itself may not be available, so > separate fallback code is necessary regardless. Surely you just need an OpenCL library present? It would be like falling back on Mesa if hardware GL is not available: the li

Re: [Bf-committers] Blender and OpenCL

2010-08-28 Thread Ralph Giles
2010/8/28 Arturo José Pérez Verdú : > #if defined(__APPLE__) || defined(__MACOSX) > #include > #else > #include > #endif Yes, sadly this bit of incompatibility was willfully propagated from OpenGL. :( -r ___ Bf-committers mailing list Bf-committers@

Re: [Bf-committers] Blender and OpenCL

2010-08-28 Thread Arturo José Pérez Verdú
Hi again, El 28/08/2010, a las 21:50, Ralph Giles escribió: > 2010/8/25 Arturo José Pérez Verdú : > >> When you are talking about the fallback to normal CPU execution you are >> talking about an optimized CPU execution? I mean, as I understand OpenCL, >> there's always at least one OpenCL capa

Re: [Bf-committers] Blender and OpenCL

2010-08-28 Thread Ralph Giles
2010/8/25 Arturo José Pérez Verdú : > When you are talking about the fallback to normal CPU execution you are > talking about an optimized CPU execution? I mean, as I understand OpenCL, > there's always at least one OpenCL capable device, the CPU. This depends on the implementation. OpenCL code

Re: [Bf-committers] Blender and OpenCL

2010-08-28 Thread Lukas Tönne
I have tried out your patch, nice work :) Here are some more thoughts on how to process data in the node tree. I hope i'm not getting too verbose or tell you guys obvious stuff ;) Basically when talking about data in the tree i see two different types of dependency: 1. Inter-node dependency ("ver

Re: [Bf-committers] Blender and OpenCL

2010-08-27 Thread Jeroen Bakker
Hi, I have done a proof of concept on Blender and OpenCL. My victim is the compositing node system. I learned a lot during these 4 hours and put it on my website including the patch. http://sicg.atmind.nl/index.php?option=com_content&view=article&id=21:blender-compositor-and-opencl&catid=8:blen

Re: [Bf-committers] Blender and OpenCL

2010-08-25 Thread François T .
any how, all this sounds good to my ears ! if it gets implemented for nodes (especially compositing), count me for testing :) 2010/8/25 Jeroen Bakker > > > On 08/24/2010 07:46 PM, Xavier Thomas wrote: > > You might be interested in MiniCL used in the bullet physics library. > > It permit to use

Re: [Bf-committers] Blender and OpenCL

2010-08-25 Thread Jeroen Bakker
On 08/25/2010 07:36 AM, Lukas Tönne wrote: > You may have noticed the new simulation nodes i am currently working > on ( http://phonybone.planetblender.org/ ). Until now i didn't really > pay attention to OpenCL, in fact i even made my own multithreaded > batch processing system for object data lik

Re: [Bf-committers] Blender and OpenCL

2010-08-25 Thread Jeroen Bakker
On 08/25/2010 01:07 PM, Arturo José Pérez Verdú wrote: > I'm also playing with OpenCL and I think is a good thing to start exploring. > > I started playing with it recently and I consider that my knowledge on this > area is still small but I will follow your discussion very closely. > > I also wan

Re: [Bf-committers] Blender and OpenCL

2010-08-25 Thread Arturo José Pérez Verdú
I'm also playing with OpenCL and I think is a good thing to start exploring. I started playing with it recently and I consider that my knowledge on this area is still small but I will follow your discussion very closely. I also want to make some observations to what has already been said. Most o

Re: [Bf-committers] Blender and OpenCL

2010-08-24 Thread Jeroen Bakker
On 08/24/2010 07:46 PM, Xavier Thomas wrote: > You might be interested in MiniCL used in the bullet physics library. > It permit to use OpenCL on CPU only. > > Thanks, but I have a question on MiniCL as I am not sure if this is suitable. IMO you need different code for CPU's and GPU's. CPU

Re: [Bf-committers] Blender and OpenCL

2010-08-24 Thread Lukas Tönne
You may have noticed the new simulation nodes i am currently working on ( http://phonybone.planetblender.org/ ). Until now i didn't really pay attention to OpenCL, in fact i even made my own multithreaded batch processing system for object data like vertices and particles. Yesterday i took a closer

Re: [Bf-committers] Blender and OpenCL

2010-08-24 Thread Thomas Dinges
Hi, this is a very good idea and it would be awesome to have OpenCL for Fluids in Blender! Thomas Am 24.08.2010 19:18, schrieb Jeroen Bakker: > Hi all > > I have been experimenting with OpenCL and are planning a basic framework > to support it in Blender. > > main features are: >* OpenCL i

Re: [Bf-committers] Blender and OpenCL

2010-08-24 Thread Reuben Martin
Yo, back on Tuesday, August 24, 2010 Keir Mierle was all like: > On Tue, Aug 24, 2010 at 10:18 AM, Jeroen Bakker wrote: > > > I have been experimenting with OpenCL and are planning a basic framework > > to support it in Blender. > > Can't the default be "auto" so that OpenCL is used if present,

Re: [Bf-committers] Blender and OpenCL

2010-08-24 Thread Xavier Thomas
You might be interested in MiniCL used in the bullet physics library. It permit to use OpenCL on CPU only. 2010/8/24 Jeroen Bakker : > Hi all > > I have been experimenting with OpenCL and are planning a basic framework > to support it in Blender. > > main features are: >  * OpenCL is disabled by d

Re: [Bf-committers] Blender and OpenCL

2010-08-24 Thread Ian Johnson
Hi Jeroen, I have been experimenting with OpenCL in Blender, trying to add an accelerated particle system to the BGE. I've been keeping track of my progress on my blog: http://enja.org I have some code up on github: http://github.com/enjalot/EnjaParticles and a fork of Blender with some modificati

Re: [Bf-committers] Blender and OpenCL

2010-08-24 Thread Keir Mierle
On Tue, Aug 24, 2010 at 10:18 AM, Jeroen Bakker wrote: > Hi all > > I have been experimenting with OpenCL and are planning a basic framework > to support it in Blender. > > main features are: > * OpenCL is disabled by default, CPU fall-back must ALWAYS be > available. OpenCL can be enabled with

[Bf-committers] Blender and OpenCL

2010-08-24 Thread Jeroen Bakker
Hi all I have been experimenting with OpenCL and are planning a basic framework to support it in Blender. main features are: * OpenCL is disabled by default, CPU fall-back must ALWAYS be available. OpenCL can be enabled with command-line parameter * Compiler directive to completely disable