Workstation for Quantum Espresso | Information dissemination media for research and development TEGAKARI

A customer involved in ceramics research and development asked us about a workstation for first-principles calculations using Quantum Espresso.
We would like to have an optimal configuration for a budget of around 200 million yen.

The workstation that will be introduced will be used exclusively for first-principles calculations, and it is assumed that the calculation target will be around 100 atoms, with a maximum of around 200 atoms.

Also, when running Quantum Espresso on another machine, it used about 100GB of memory, so we requested a memory capacity of 128GB or more.
I hear that your preferred OS is Windows, and that you would like to run Quantum Espresso on WSL.

Based on the information we received from our customers, we proposed the following configuration.

CPU	Intel Xeon Gold 6438Y+ 2.00GHz(TB 4.00GHz) 32C/64T × 2
memory	256GB REG ECC
storage	1TB M.2 SSD
Video	NVIDIA T400 4GB (MiniDisplayPort x3)
network	on board (10GBase-T x2)
Housing + power supply	Tower type housing + 1000W
OS	Microsoft Windows 11 Professional 64bit
Other	HyperThreadring disabled

Quantum EspressoWhat are the machine specs required for operation?

It is a 4-CPU configuration that uses 2th generation Xeon Scalable.
There are no clear operating requirements for Quantum Espresso.
Therefore, although it is possible to operate it without a high-end computer, the performance during operation depends on the specifications of the machine.

Quantum Espresso generally places emphasis on parallel processing performance, such as the number of CPU cores and GPU, and requires memory capacity accordingly.
In this configuration, we place emphasis on the CPU's parallel processing performance, and select a CPU with emphasis on the number of cores within your budget.

Intel Xeon Gold 6438Y x2 is equipped with a total of 64 cores, and high parallel performance can be expected.

In addition, the current calculations are consuming 100GB of memory, so in consideration of the possibility that the calculation scale will increase in the future, we have installed 2GB, twice the memory capacity you specified. .

Although Quantum Espresso has a GPGPU option implemented, it is basically assumed that double-precision arithmetic (fp64) will be used.
GPUs with FP64 performance are extremely expensive, and it was extremely difficult to fit them within the budget for this project.
Therefore, in our proposal, we prioritized staying within the budget, and selected the NVIDIA T400 4GB video card, which is an entry-class workstation model that only takes into account drawing without considering the GPGPU option.

WSLHow Quantum Espresso works on

Please note the following when using Quantum Espresso on WSL:More than 64 threads may not be recognized correctly on WSLWe are confirming the information.

reference:WSL 2 uses half the number of cores on AMD Threadripper 3990X

This configuration has a total of 2 CPUs with 64 cores/128 threads, so there is a possibility that only half of these will be recognized on WSL.
In order to avoid the above problem, we are assuming that HyperThreading will be disabled in the CPU settings at the time of shipment and will be used with 64 cores/64 threads.
This is expected to allow all 64 physical cores to be used on WSL.

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.

Keyword

・What is Quantum Espresso?
Quantum Espresso is open source software written in Fortran95 and distributed free of charge under the GNU General Public License. First-principles electronic state calculation program. It is a first-principles calculation code based on density functional theory. It is capable of optimizing atomic positions and cell structures, calculating electronic states and band structures, analyzing phonon vibrations, and calculating neutron scattering. It operates on multiple platforms and supports parallelization using MPI and OpenMP. It is widely used for theoretical analysis of material properties and the search for new substances, and is also a useful tool in computational biology.