Hi Tat,
That is explained in the paper on bioRxiv. Basically, each body is like a
reference in 3D auto-refine, but due to overlapping masks more reference
maps (but not the more expensive weighted sums and sums of weights) need
to be kept in memory. We did all our calculations with 3 or 4 bodies using
GPU workstations with 256GB of RAM.
HTH,
Sjors
> Hi all,
>
> Can't wait to try on the multi-body refinement on Relion3. In the process
> of setting up hardware for cryoEM computation, I wonder how the memory
> usage scales with increasing bodies and/or overlaps? Thanks a lot.
>
> Best,
> Tat
>
--
Sjors Scheres
MRC Laboratory of Molecular Biology
Francis Crick Avenue, Cambridge Biomedical Campus
Cambridge CB2 0QH, U.K.
tel: +44 (0)1223 267061
http://www2.mrc-lmb.cam.ac.uk/groups/scheres
|