RELION (REgularised LIkelihood OptimisatioN) is a stand-alone computer program for the refinement of 3D reconstructions or 2D class averages in electron cryo-microscopy.
RELION is available on the Owens cluster. The versions currently available at OSC are:
| Version | Owens | Pitzer | Ascend | Note |
|---|---|---|---|---|
| 3.1-cuda10.1 | X | Built with CUDA 10.1 | ||
| 4.0-cuda10.1 | X | RELION 4.0 beta2 Built with CUDA 10.1 |
||
| 4.0.1 | X | X | X | Built with CUDA 10.2 (11.8 for Ascend) |
| 5.0b | X | Built with CUDA 11.8 |
You can use module spider relion to view available modules for a given machine. Feel free to contact OSC Help if you need other versions for your work.
| Cluster | RELION | CTFFIND | MotionCor2 | GCTF | ResMap | Unblur & Summovie |
|---|---|---|---|---|---|---|
| Pitzer | 4.0.1, 5.0b | 4.1.14 | 1.4.4 | 1.18 | 1.1.4 | 1.0.2 |
| Ascend | 4.0.1 | 4.1.14 | 1.4.5 | 1.18 | 1.1.4 | 1.0.2 |
RELION is available to all OSC users.
MRC Lab of Molecular Biology, Open source
To set up the environment for RELION on the Owens cluster, use the command:
module load relion/version
where version is chosen from the available versions (omitting the version will load the default version).
To set up the environment for RELION on the Pitzer cluster, use the command:
module load relion/version
where version is chosen from the available versions (omitting the version will load the default version).
Hybrid MPI+OpenMP jobs utilizing programs built with the Intel compiler and MVAPICH2 may experience hangs when running on multiple nodes. This issue is attributed to a known problem in MVAPICH2 built with the Intel compiler stack.
All versions of Intel+MVAPICH2 have been removed to address this issue.
RELION versions prior to 5 may exhibit poor performance in hybrid MPI+OpenMP jobs when the number of MPI tasks exceeds 4 on multiple nodes.
If possible, limit the number of MPI tasks to 4 or less for optimal performance. Consider using RELION version 5 or later, as newer versions may include optimizations and improvements that address this performance issue.