Greetings. I am contemplating buying a desktop with the following specs to enhance the performance of molecular dynamic simulations (GROMACS) and other computational works (e.g., homology modeling, energy optimization using Gaussian, etc.). Could this desktop significantly improve the MD and reduce the simulation time? Using the current laptop, I needed around 5 days to run a 30 ns simulation, which is extremely slow.
I7-14thgen
32GB ram
1TB SSD
RTX 3050
Windows 11 pro
Hello,Dear
Chun Hoe Tan.
Your System components are sufficent to run samall molecules calculation..while when you are dealing with big molecules such as protein and polymers you need server becz these calculation needs more number of nodes..higer the numbers of Node,less will be the time,also depends upon number of processor per node....generally 8-10 processor/node are offerd on server.
hope the this are cleard now.
Regards
Vikrant Singh
Dear Vikrant Singh
Thank you for your reply. Where do I get the server?
It depends on what you want to run, specifically how big is your box and what are the time scales you are aiming for.
NVIDIA GPUs are a must nowadays, especially in workstations, if you want to run MD simulations. Here is where you will spend the most, and you should aim for the best you can get with the budget. RTX 4090 would be ideal, but also the previous generation like the RTX 3080 are still quite good.
I would avoid intel but get some 16core Ryzen, and also I would avoid Windows and put some unix system (Ubuntu 24 for example). This will save you also some money - Ryzen are usually less expensive and you don't have to pay the windows license. Also, in my experience some brands tend to 'personalize' the computer with additional software, especially on the motherboard/BIOS level, like some Alienwares. I had a ton of problems with them to have a working Ubuntu distribution, and usually they mount the Intel chipsets that have some kind of fake cores split as high/low performance. These are usually just a pain, try to avoid them if you can.
Memory wise, for RAM not less than 32GB, possibly 64GB (MD uses basically no RAM but the analysis will be on RAM). SSD with min 2TB as well to have some kind of storage, or less but with external storages ready.
I have a AMD Ryzen 9 7950X 16-Core Processor with RTX 4080 and run 50ns/day circa for systems with 650000 atoms. Not bad at all.
With the few details you post about the workstation, and the absence of details on the laptop, make it hard to give general answers but given the numbers you post then most likely the workstation will be faster than that. If fast enough, well, that depends on your systems and expectations!
Dear Nicola Piasentin
Greetings, thank you for your comprehensive reply. I found a laptop with the spec you mentioned. Do you think I could opt for a gaming laptop rather than a desktop? I wanted a gaming laptop for my computational work, but I am quite worried about overheating if I continuously run MD with a laptop. A desktop won't have this overheating issue, but it is bulky and space-consuming.
In terms of MD, this is a no brainer: a desktop.
In general, you have much better heat management, better power, and the costs are lower for the same pieces. Consider also that the desktop is upgradable, so if in a couple of years you feel like you need more ram, a better gpu, etc, you can change those on a desktop, while on a laptop in general or you can't or it's not worth the money. Heat in particular is going to put a TON of stress on the laptop. For power, heat, and volume constraints laptops are also generally equipped with weaker hardware (e.g. dowgraded GPUs).
Clearly then it is up to you, as a laptop can be moved easily and consumes less power. However, I would never count on it for long runs, and I wouldn't buy a laptop if the main usage of the machine is MD sims.
Chun Hoe Tan here i pin some suppercomputing server links you may contact them and will utilize the server.
IIT-BHU Param_Shivya (https://iitbhu.ac.in/cf/scc )
IIT-Delhi PADUM ( https://supercomputing.iitd.ac.in )
Yes, the desktop with an i7-14th gen, 32GB RAM, RTX 3050, and SSD will significantly improve performance for molecular dynamics simulations (GROMACS) compared to a laptop. The RTX 3050 will speed up simulations, especially for GPU-accelerated tasks, and the 32GB RAM and SSD will help with faster data handling and memory management. However, for even better performance, consider a higher-end GPU (e.g., RTX 3060 or better) if budget allows.
all of which can we easily done at mdsim360.com, a new platform that lets you run MD simulations entirely online without local installation.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Computer Graphics and Animation
- 3D