END OF AN ERA, FRACTALFORUMS.COM IS CONTINUED ON FRACTALFORUMS.ORG

hello!

i would like to use mandelbulber with opencl support.
opencl works with my ati radeon 5850 card (catalyst 13.4)
i'm using debian (unstable, /mandelbulber_1%3a1.15-1_i386.deb) and the debian version comes without opencl-support, so i tried to compile mandelbulber from svn (rev 348) but failed:

...
Invoking: GCC C++ Linker
g++ -o"mandelbulber" -L/usr/lib/x86_64-linux-gnu/  ./src/Render3D.o ./src/algebra.o ./src/callbacks.o ./src/cimage.o ./src/cl_support.o ./src/common_math.o ./src/database.o ./src/files.o ./src/fractal.o ./src/image.o ./src/interface.o ./src/morph.o ./src/primitives.o ./src/settings.o ./src/shaders.o ./src/texture.o ./src/timeline.o ./src/undo.o   -lpng -ljpeg  `pkg-config --libs gtk+-2.0 gthread-2.0`
./src/Render3D.o: In function `Render(sParamRender, cImage*, _GtkWidget*)':
Render3D.cpp:(.text+0x3389): undefined reference to `netRender'
Render3D.cpp:(.text+0x34a5): undefined reference to `CNetRender::sendDataToClient(void*, unsigned int, char*, int, int)'
...


any suggestions, what i am doing wrong?
and: is the svn-version of mandelbulber the correct one for opencl-support?


thanks for tips & hints
d13b

Zero megabytes seems a bit odd, but further down it's 512 MB of "global" memory on the graphics card. A limit of 128MB for an individual allocation is normal for this generation of AMD GPU, too, as far as I know.

And while I don't know mandelbulber specifically, I think you are guessing right that each individual formula needs to be prepared for OpenCL by a programmer. It's not as simple as flicking a switch.

to get opencl working in debian linux, i installed the debian amd-opencl-dev package amd's proprietary drivers (amd-driver-installer-catalyst-13-4-linux-x86.x86_64.zip and AMD-APP-SDK-v2.8-RC-lnx32.tgz from amd.com), then configured X to use the fglrx drivers. (aticonfig worked).

the bigger task was to compile mandelbulber, since make didn't work out and i had to rewrite some make helperfiles and enable the opencl support in the sources (#define CLSUPPORT was commented in cl_support.hpp and cl_support.cpp)

seems i will have to wait until opencl-support is more complete in mandelbulber, but it's indeed much faster than cpu-rendering from what i can tell now.

best
d13b