We have been experimenting with image based lighting and importance sampling of a HDRI maps - as well as unifying our various ray trace related techniques into a single framework (codename: Cyclops). This framework is to be used for some of our future ray trace related projects, e.g. quick rebuild of acceleration structure as well as predictive rendering.
... oh, and everything is GPU based - coded from the ground up (Cuda and OpenCL).
#1 by Mikael Christensen on November 24, 2010 - 12:32 pm
Quote
So, how fast is this? What is the rendering time for an image such as the one above?
#2 by krys on December 4, 2010 - 12:46 pm
Quote
Hi,
Have you mixed cuda and opencl ?!!
Or you have write your renderer in cuda and another in opencl ?
We are also working on an OpenCL framework/api for rendering. Maybe we can collaborate ?
#3 by admin on December 27, 2010 - 7:09 am
Quote
We will return in the new year with speed measurements
#4 by admin on December 27, 2010 - 7:13 am
Quote
The CUDA and OpenCl pipelines's are not directly mixed. We are always interested in collaboration, let me get back to you in the new year - perhaps you could give us some more details on your work ?
#5 by Krys on January 19, 2011 - 3:24 pm
Quote
Please contact me through my email.... I will give you more details about our work.
Regards