[Bf-committers] Blender and OpenCL
Arturo José Pérez Verdú
arturjosep at gmail.com
Sun Aug 29 01:04:35 CEST 2010
El 28/08/2010, a las 21:50, Ralph Giles escribió:
> 2010/8/25 Arturo José Pérez Verdú <arturjosep at gmail.com>:
>> 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 *can* be compiled to
> run on the CPU, and for example Apple's implementation supports this,
> but I believe some of the GPU vendor's implementations do not; that
> is, they can only compile targetting the GPU, not the CPU.
> But as Jeroen points out, OpenCL itself may not be available, so
> separate fallback code is necessary regardless.
Yes, what a pity. :-( I'm using OpenCL on MacOSX and never cared about this.
So in this case extra maintenance cost must be assumed.
By the way, I haven't tried the patch yet but just giving a look I think we will need this to compile under MacOSX
#if defined(__APPLE__) || defined(__MACOSX)
The headers are different in MaxOSX.
More information about the Bf-committers