Dear Community and Developers,
I’m currently in search for the optimal hardware setup for calculations using CryoSPARC. In the documentation there are some benchmarks for the homogeneous and local refinements: e.g. for 1200 box - 24GB of VRAM, but how much VRAM would be needed to run smoothly heterogeneous refinement or 3D Flex job of similar box size? How can be that optimized in order to be able to still run these jobs with less VRAM?
Thank you in advance,
Valentyn