We know that PhotoScan is quite hungry for RAM, which makes processing of larger projects as well as higher quality processing quite cumbersome on "normal" computers. Therefore, it would be great if Photoscan could automatically split the project into smaller pieces based on the available RAM.
This would allow to process bigger projects without investing thousands of *insert your currency here* in hardware. It would also be much easier than manually generating smaller chunks in a large project and at the end merging them, which people are doing right now as a workaround.
Cheers
SAV