I've updated to the latest version and gave it another try.
I deactivated one gpu, the other gpu and activated both and cpu.
None of those tests where positive. At some point Metashape stopped for hours and did nothing.
It was also not reacting to the cancel button.
Normal setup:
CPU1: Socket E5 2683 V3
CPU2: Socket E5 2683 V3
GPU1: Vega 64 ROG Strix
GPU2: Vega 64 ROG Strix (Crossfire not active)
RAM: 128 GB
Windows 10 64 bit
SSD Free space C: 194 GB (Windows)
SSD Free space E: 199 GB (project)
Both GPUs are activated. CPU is disabled.
Image set:
- this time 1129 fisheye cameras aligned
- alignment: highest, not generic, reference(source), 40000, 0, adaptive camera model fitting
- a lot of overlapping images
- therefore disabled half of the images
- mesh via depth maps: high, interpolation enabled, mild, no vertex colors
I also noticed that it could have to do with the optimization process.
- python script optimizes 10 times for me
- f, cx, cy, k1, k2, k3, p1, p2
- Do you have any suggestions on which parameters should be activated or disabled?
When estimating camera locations GPU1 is active at about 25% (attachment) even thought bth GPUs are activated.
Most problems occur when calculating depth maps. Metashape just stops after some time. CPU & GPU usage is then 0%.
CPU was running 100%, both GPUs where working.
The same system was able to calculate models with double the images even at ultra quality before.
Thank you!