Dear Members,
I am planning to buy a PC with the following configuration-
Processor- 10th Gen Intel® Core™ i9-10900 (20 MB cache, 10 cores, 20 threads, 2.80 to 5.20 GHz Turbo, 65W)
Memory- 32 GB DDR4 RAM
Video Card- AMD® Radeon™ RX 640 4 GB GDDR5 low profile, 2 mDP, DP
Hard Drive- 512 GB, M.2 2230, PCIe NVMe, SSD, Class 35, 2.5 inch 2TB 5400rpm Hard Disk Drive
This configuration is good for my current work which involves RNA-Seq analysis and protein-ligand docking. But, if I want to do molecular dynamic simulation later (in Gromacs), would AMD® Radeon™ RX 640 4 GB card would be sufficient? I probably need to simulate a protein-ligand complex for up to 100ns. My protein of interest has approximately 500 amino acids.
Please provide your valuable suggestions. I would also welcome suggestions on an alternative PC configuration to improve performance without a drastic increase in price.
Dear S.K. Sanu
It depends a lot on the type of simulations you want to run. If you go for an AA description, then a solvated box will be easily at least of the order of tens of thousands of atoms.
My two cents are that you should build and benchmark a bit your system before spending any money. In this way you will get an idea of the weight (how many atoms? Which forcefield? Are 100ns sufficient?) of the sims. Is there any way you can have access to a cluster to test this out?
The GPU of the configuration seems a bit weak, but prices for GPU (and everything else tbh) are madness right now. Also avoid laptops, heat dissipation is a huge problem when running sims for days and can negatively impact on the longevity of your machine.
Maybe if the simulations are relatively short you could get more value for your money if you spent a bit on buying some computational time through your research/industry institution rather than a new machine?
Best,
Nicola
- Badges
- Science topic