FAQ 002342 CS


What are the optimal hardware requirements for RSTAB and RFEM?


​In general, it is difficult to make a generalized statement regarding optimal hardware. It always depends on the models you wish to calculate. The following information may be useful for configuring a computer optimized for your purposes.


Depending on the individual application, the system for RFEM can be optimized in two different ways:

1. As many cores as possible.
2. Fewer cores but the highest possible turbo boost.

The first alternative is reasonable under the following conditions:

- Large models with solids and / or many surface elements
- Relatively few load cases
- Extensive calculations with RF-STABILITY and RF-DYNAM-Pro

The more CPU cores there are in a system, the greater the overhead for communication between them becomes. More than 10 cores are therefore not much of a performance increase. Hence, it makes little sense to use systems with several processor sockets.

The second alternative is reasonable in the following use cases:

- Member models
- A lot of load cases
- Extensive calculations with modules (except RF-STABILITY and RF-DYNAM-Pro)

Which of the alternatives suits you better can be tested in the following manner: Start a typical calculation. Keep an eye on the processor load in the task manager. If it is at ~100% for processors without hyperthreading or at ~50% for processors without hyperthreading, choose alternative 1.

If the processor load is at least at 25% (processors without HT) or 12.5 (processors with HT), alternative 2 is the more suitable one.
For alternative 1, we recommend one of the following processors:

- Intel® Core™ i7-8700K
- Intel® Core™ i9-8950HK
- Intel® Core™ i9-7900X
- Intel® Core™ i7-7820X

For alternative 2, we recommend one of the following processors:

- Intel® Core™ i7-8700K
- Intel® Core™ i9-8950HK
- Intel® Core™ i7-8086K
- Intel® Core™ i5-8600K
- Intel® Core™ i7-7740X

RFEM does not benefit from hyperthreading. There are therefore no disadvantages to not having a processor with hyperthreading.

RSTAB can only use one processor core. If a computer is to be used only for RSTAB, equipping it according to alternative 2 makes sense.
Generally, Intel processors are recommended. We do not recommend using AMD processors, but have no experience or feedback regarding Ryzen processors.

Modern processors have sophisticated temperature management. Depending on the processor temperature, the clock speed is adjusted dynamically, which means that good cooling can increase a processor's performance. We therefore recommend a generously sized computer case and a good cooling fan for the processor.

XEON processors do not have any significant advantages in regards to the computing speed. In these processors, the same cores as in the Core i processors are used. The additional functions these processors provide do not lead to a performance increase in RSTAB and RFEM.


A quick RAM connection is important for the speed, especially for RSTAB. For that reason, DDR4 RAM should generally be used.
The optimal RAM size depends on the FE mesh of the model. Unfortunately, the size of the required memory cannot directly be inferred from the number of FE nodes. Testing it is the only way.

For that, use a model from your practical work that is as large as possible. Select the direct solver (see Figure 1).

Keep an eye on the utilization of the working memory in the task manager. If it approaches 100%, your computer does not have enough RAM.

If the RAM size is sufficient for the calculation, adding more memory will not result in increased speed.

Sometimes it makes sense to perform comparative calculations of different variants parallelly in several instances of RFEM. For example, if you wish to run 2 parallel instances of RFEM, the required working memory doubles.

Equipping the mainboard with RAM in such a way that it is easily upgradable is sensible.

ECC memory, as it is offered in some workstations, has no practical use for RFEM in terms of speed.

Graphics card

For the video card, we recommend a graphics chip from Nvidia. The integrated graphics processors from Intel are unsuitable.

We do not recommend the expensive Quadro cards and instead recommend a "gaming card" of the GTX series with Pascal architecture (GTX 10XX). In terms of hardware, GTX and Quadro cards differ only slightly. The drivers for the GTX cards are usually less of a problem with regards to RSTAB / RFEM and the GTX cards are less expensive than the Quadro cards.

The size of the video memory of the card is irrelevant for RSTAB / RFEM.

What is important is that the latest driver for the graphics card is installed.


Regarding computing speed, the use of an SSD instead of an HDD is advantageous only in exceptional cases. They do have an advantage when it comes to saving or reading large RFEM files, though.


The software that is running in the background may be pivotal for the calculation speed.

Antivirus software running in the background may significantly slow calculation. During calculation, RFEM has to write many files into the work folder and read from it again. It could have a positive impact on the computing speed to remove this folder from the real-time monitoring of your antivirus software (see Figure 2).

Klíčová slova

hardware, CPU, processor, RAM, SSD, graphics card, system requirement



Nenalezli jste odpověď na Vaši otázku?
Kontaktujte nás nebo nám zašlete Váš dotaz prostřednictvím online formuláře.

+420 227 203 203


RFEM Hlavní program
RFEM 5.xx

Hlavní program

Program RFEM pro statické výpočty metodou konečných prvků umožňuje rychlé a snadné modelování konstrukcí, které se skládají z prutů, desek, stěn, skořepin a těles. Pro následná posouzení jsou k dispozici přídavné moduly, které zohledňují specifické vlastnosti materiálů a podmínky uvedené v normách.

Cena za první licenci
3 540,00 USD
RSTAB Hlavní program
RSTAB 8.xx

Hlavní program

Program pro statický výpočet a navrhování prutových a příhradových konstrukcí, provedení lineárních a nelineárních výpočtů vnitřních sil, deformací a podporových reakcí.

Cena za první licenci
2 550,00 USD