few minutes ago iam shock about compiling a 'simple' map for warehouse demo takes arround 10 minute to compile in my quadcore i7 965,is there plan to accelerating compiling map time by cuda?is there any plan of these in conitec 'todo' list?i've felt that compiling map somewhat pain in the ass,for ati there is a alternative for cuda which is called ati stream,these 2 technology have different sdk so it's mean it would be 2 exe for nvidia cuda dan ati stream.if there is a plan for this it would be great!!!
i know that porting current base code into parallel computing would be a huge task but i and the others? hope for these.not to mention currently yes only 'big' app supported cuda,3dsmax,adobe cs families,maya,matlab,etc but it would huge marketing boost for 3dgs.

well not only map compiler,it will be good if the engine ported to cuda also

example app performance boost ported to cuda!! : http://www.youtube.com/watch?v=tOKLo8zSWbY

List of supported Cuda enabled gfx : http://www.nvidia.com/object/cuda_learn_products.html

STL like library for cuda : http://code.google.com/p/komrade/

cheers,

i would certainly like a CUDA Accelerated Map Compiler, but I think JCL is going to jump in and go for his routine he has been following the last few years and say "PUT MOAR RAM IN YOUR COMPUT0RZ!"

No, following my routine I usually say "LORD, THROW MOAR BRAIN FROM HEAVEN!".

A CUDA accelerated map compiler is certainly nice to have, and we might have it one day, but a good alternative at the moment is not to set up the map compiler settings in a way that it needs 10 minutes for your warehouse.

oh well,eventough i put 32 GB ddr3 in my pc would be useless because the engine itself it's not 64bit.


ay ay boss