@Jonathan: you can share the repro steps in your e-mail exchange with Zuzana
Yes, I can try, but this is really very straightforward :
-add my pictures (from several folders)
-align
-clear the automatic reconstruction region
-reconstruct in D3
I haven’t got too much experience with XMPs, but the main use is that you export the camera parameters of an alignment for each image individually. If present, RC will by standard read those files and use it for the alignment, skipping all calculations. So in theory you wouldn’t even need to work with components. But again, this is lukewarm knowledge from a while back…
mmm ok thanks for the extra infos Götz.
I can’t be the only struggling on reconstructions on last version? I have the same issue (memory leak and swap at the end of processing) in nearly all my projects, on different datasets and on my 3 machines (1 CLI license and 2 promos) :
-my big tower i7 3x1080Ti 128GB RAM
-my big laptop i7 1x1070 64GB RAM
-my mini on the field laptop i5 1x740M 12GB RAM
Anyone?
Can we allocate less memory to RC so that when it tries to get more it does not go into virtual memory?
Is there something to do?