Used several renders to find most efficient size for GPU and CPU, had a problem when i set the priority to realtime and lost my record for most of my tests. here are the ones i managed to save.
all times are (min:sec:milisec)
Hardware: Custom Rig (designed for 3d game programming and testing)(no SLI)
Intel i7 5930K (6 core, 12 threads) 3.5GHz
64GB ram
Win 10
Gaphics
- NVidia GTX 980
- NVidia GTX 980 Ti
64x64 both GPUs 8:11:47 with progressive refine
64x64 ti gpu 9:41:27
64x64 non-ti gpu: 9:24:03
64 x64 CPU 15:06:26
128 x128 duel gpu, no opensubdividecomputer, no progressive refine
1:05:50
CPU
128 9:57:30
64 15:06:26
32 8:57:31
No Spatial Split Duel GPU no progressive render
00:52:65* best time
Time saved = Build time (13ish sec)