Since atom size plays one of the important roles while performing Molecular Dynamic Simulations. Does anybody know what should be the required RAM of the processor based on the atom size of the protein-ligand complex in Gromacs?
Getting the maximum MD simulation performance depends on the system benchmark. Therefore 1) the number of physical cores 2) cache 3) clock speed 4) types of hard-disks, 5)GPU, 6) program configuration 7)ForceField all are important factors.
You can read these articles to get an idea about benchmarking on MDS in Gromacs.
Article MDBenchmark: A toolkit to optimize the performance of molecu...
Pro-Tips: If your system size is nearly 60k to 70K atoms, then in a laptop having 8 cores, 8 GB/16 GB RAM with NVIDIA GPU (PASCAL or TURING CUDA11.1 enabled), and a properly configured Gromacs, you can achieve 50 ns+ /day simulation. However, a good workstation is always recommended for an MDS job for more than 100K atoms. For millions of atoms, it definitely required a computer cluster having multiple nodes for parallel calculations.
For performing MD simulation, 32GB of memory will be enough with more than 16cores. Even you can run on a lower configuration system. But what is very important for MD simulation is having GPU that will accelerate the overall job and even you can completely shift your job from CPU to GPU. Also Desmond which is available with Schrodinger's Maestro, only supports GPU.
Performing MD stimulation, 16-32 GB of memory will sufficient although you run on a lower grade system. Clock Speed - 3.7 (Min) run well & Minimum i3/ ryzen 3 processor suggests due to heat generated during program running. I usually did my MD stimulation on Gromacs. GPU is an important factor for simulation I recommend NVIDIA (at least 2GB 980gtx).
When doing MD stimulation, 16-32 GB of RAM will enough, even if the machine is of a lesser grade. Clock Speed - 3.7 (Min) operate smoothly, and Minimum i3/ryzen 3 processor advises due to the heat created when operation. But what's really significant for MD simulation is having GPU that will speed the total process and even you may switch your task from CPU to GPU. Desmond, accessible with Schrodinger's Maestro, only supports GPU.