Parallel export/ export performance

Hi,

For whatever reason C1 exports only 1 picture at a time on a 4 core machine. The CPU is idling at 30 - 50%, the GPU just 18 - 19%. At the same time, memory is used 79% (of 16GB), with the IMGCORE process hitting loads of page faults.

I do not understand what is going on - one picture is about 50MB on disk, so how can you possibly use a data structure inefficient enough to fill 16GB? In case IMGCORE is doing all the processing, it would probably require less than an afternoon's work to spawn multiple instances in parallel (in case you do not want to move to a "real" multi-threaded implementation). 

Is there any "parallelization setting" I missed? Otherwise: Could someone look into this? There seems to be an extremely quick win opportunity to dramatically improve export performance.