VMD-L Mailing List

From: Vermaas, Josh (vermaasj_at_msu.edu)
Date: Mon Jan 16 2023 - 09:22:38 CST

Hi Hocine,

If its 10GB of dcd, and you want to be able to load it all into RAM at once, you'd need more than 10GB of RAM, + whatever you need to represent the structure in VMD's data structures. Since the operating system also wants some RAM, your 12GB of RAM is awfully close to the minimum specs. However, unless you *know* that you need all the frames, just load your trajectory with a stride of 2, and that will half the memory requirement, since VMD will then only load every other frame.

-Josh

On 1/16/23 08:27, Hocine el khaoudi enyoury wrote:
Dear VMD-users

Does anyone know what would be the amount of ram and/or gpu needed/recommended, for the analysis and visualization of long trajectories? Is there any kind of rule of thumb? For example, for 10 gb of data, what would be recommended?

This is the equipment that I got right now:
Multithreading available, 12 CPUs detected.
Info) Free system memory: 12742MB (81%)
Info) Creating CUDA device pool and initializing hardware...
Info) Detected 2 available CUDA accelerators:
Info) [0] GeForce GTX 1060 6GB 10 SM_6.1 @ 1.76 GHz, 5.9GB RAM, AE2, ZCP
Info) [1] GeForce GT 740 2 SM_3.0 @ 0.99 GHz, 2.0GB RAM, KTO, AE1, ZCP


Thanks in advance

Best Regards,

Hocine


--
Josh Vermaas
vermaasj_at_msu.edu<mailto:vermaasj_at_msu.edu>
Assistant Professor, Plant Research Laboratory and Biochemistry and Molecular Biology
Michigan State University
vermaaslab.github.io