Machine for electromagnetic field simulation | Information dissemination media for research and development TEGAKARI

A customer who is considering introducing electromagnetic field simulation software (CST-Studio) consulted us about introducing a PC. Since CST-Studio adopts FIT & PBA mesh separation, it is possible to shorten the processing time by GPU, so we would like you to propose an appropriate configuration.
As a specific request, we are asking for the following conditions.

・CPU: Xeon
・Memory: 256GB to 512GB (I want to secure four times the GPU memory)
・GPU: NVIDIA product with high memory capacity and high performance
・OS: Windows
・Budget: About 150 million yen

Based on the conditions you contacted us, we proposed the following configuration.

CPU	Xeon W5-2465X (3.10GHz 16 cores)
memory	256GB REG ECC
storage	1TB SSD S-ATA
video	Nvidia RTX A5500
network	on board (2.5GbE x1 /10GbE x1)
Housing + power supply	Middle tower case + 1000W
OS	Microsoft Windows 11 Professional 64bit

The operating requirements for CST Studio Suite are posted on the manufacturer's website below.

reference:CST Studio Suite & Opera Recommended Hardware (Dassault Systems) *Jump to external page

The following PDF is available as a GPU selection guide.

reference:CST Studio Suite 2023 GPU Computing Guide (Dassault Systems) *Jump to external page

Checking the reference page above, some of the CST Studio Suite's GPU acceleration utilizes FP64, and GPUs with low FP64 performance are annotated.Although this annotation is not unusable, it is presumed to be more of a warning that the expected performance may not be obtained depending on the application.
The configuration of this example was selected based on the above information, and it is a GPU that does not have such an annotation even on the reference page.However, support will start from software version 2022 SP5, so please check your software version when considering this.

As a recommended software requirement, a CPU with 8 to 16 cores and a base clock exceeding 3.00 GHz is said to support many applications in a well-balanced manner.Therefore, in this example, a 16-core Xeon was selected accordingly.
The video card is a high-performance product in a price range that can be realistically considered for introduction.
RTX A6000 and NVIDIA A100 are excluded from consideration because the price increase is large and it is not a realistic choice compared to your budget.

The configuration of this case study is based on the conditions given by the customer.
Please feel free to contact us even if you are considering different conditions from what is posted.

■FAQ

・What is CST Studio Suite?
CST Studio Suite is a comprehensive suite of electromagnetic field analysis software. Direct collaboration with 3D design/simulation is possible.It can be applied to a wide range of industries and applications, and is often used in the fields of EMC/EMI, antennas, and electronic equipment/semiconductors.A wide range of simulations is possible based on advanced physical modeling functions and mathematical methods.The advantage is that the calculation time can be shortened by using a high-performance dedicated mesher.

reference:CST STUDIO SUITE (Dassault Systems) *Jump to external page

・What is FIT?
FIT (Functional Integration Test) is a standard for functional safety, which is a system-level test that integrates multiple functional blocks and software modules.It evaluates not only the correct operation of individual functions, but also the functional safety as a whole.

・What is PBA?
PBA (Programmable Electronic System in Safety-Related Applications) is a functional safety standard that summarizes the requirements and methods for using electronic control systems in the field of functional safety.

・What is FP64?
FP64 is a standard for processing 64-bit double-precision floating-point numbers in hardware on GPUs.GPUs usually only support 32-bit single precision, so using FP64 will greatly improve accuracy in things like numerical calculations. The GPU requires a dedicated double-precision arithmetic unit, which is installed only in some GPUs.The computational performance is lower than that of single precision, but there are many cases where FP64 is required in the fields of scientific calculation and machine learning.