Molecular dynamics simulation machine (64 core model) | Information dissemination media for research and development TEGAKARI

A customer involved in research and development of chemical products asked us about installing a calculation machine.
The current molecular dynamics simulation machine that is currently in operation has many breakdowns, so we are considering updating it.
The budget was around 300 million yen, and they requested a configuration suitable for the above purpose.
The specs of the machine I'm currently using are as follows.

CPU: Intel Xeon Gold 6242R (3.10GHz 20 cores)
Memory: 192GB REG ECC
Storage: 4TB HDD S-ATA
Video: NVIDIA RTX A4500 20GB
OS: Windows 11 Pro for Workstations
Warranty period: 3 years

This machine is used for MD calculations and general analysis of molecular assemblies in molecular dynamics simulations, and was also intended for use in machine learning.
I also heard that the machine is often operated continuously overnight for MD calculations.

For the machine we will introduce this time, we want to shorten calculation time with a higher-performance CPU, and we will use VMD to visually confirm molecular assembly systems and use it for machine learning, so we selected the optimal video card. is the condition.

Furthermore, since quantum chemical calculations are not performed frequently, we received information that the same level of memory as the current one is sufficient.
On the other hand, the current 4TB storage capacity is not enough, so we would like an HDD of around 20TB.

Based on this information, we proposed the following configuration.

CPU	Intel Xeon Gold 6438Y+ (2.00GHz 32 cores) x2
memory	256GB REG ECC
storage	20TB HDD S-ATA
Video	NVIDIA RTX A4500 20GB
network	on board (10GBase-T x2)
Housing + power supply	Tower type housing + 1000W
OS	Microsoft Windows 11 Professional 64bit
Other	“Anshin+” 3-year take-back repair guarantee

4th generation Xeon Scalable 2 CPU configuration

It is a dual CPU configuration equipped with the latest 2024th generation Xeon Scalable series as of March 3.
With a total of 64 cores, you can expect higher performance when calculating MD compared to Intel Xeon Gold 6242R.

NVIDIA RTX A4500 20GBIs machine learning possible?

For the video card, we selected the same NVIDIA RTX A4500 20GB as the machine we are using, depending on our budget. In the system requirements of the software "VMD" that you plan to use, the video card is“Video card capable of hardware acceleration with OpenGL is highly recommended”There is only a description. NVIDIA RTX A4500 20GB is a product for mid-range workstations, and it also supports OpenGL 4.6, so it is considered to have sufficient specs for using VMD.
In addition, the performance of the video card does not have a large effect on the performance of the CPU, except for some processing such as software that supports hardware acceleration, whereas the specifications of the CPU are directly linked to the overall processing performance, so the CPU is more important than the GPU. The configuration is as follows. If you are planning to do DeepLearning in the future, please note that although NVIDIA RTX A4500 20GB can be expected to have sufficient performance for inference purposes, it may not have sufficient specs for learning purposes.

In addition, the power capacity of this machine is selected based on the current configuration, and does not take into account additional GPUs, etc.
If you are planning to add a device that consumes a lot of power, such as adding a GPU, please feel free to contact us as it is possible to change to a larger capacity power supply unit.

The configuration of this case study is based on the conditions given by the customer.
We will flexibly propose machines according to your conditions, so please feel free to contact us even if you are considering different conditions than what is listed.

■ Keywords

・What is VMD?

VMD (Visual Molecular Dynamics) is molecular modeling software developed by the Theoretical and Computational Physics Group at the University of Illinois at Urbana-Champaign. You can model, visualize, and analyze molecules using 3D graphics, and it is compatible with MacOS X, Unix, and Windows.

Reference: VMD – Visual Molecular Dynamics *Jumps to an external site

・What is Deep Learning?
DeepLearning is a type of machine learning that uses multilayer neural networks to perform advanced pattern recognition and prediction.Since it generally requires a large amount of data, it is considered an effective method when data is abundant.DeepLeanig is also widely used in fields such as image recognition, speech recognition, and natural language processing.Because it can learn complex features and relationships, it can achieve higher accuracy than traditional machine learning methods.