Operation of RFEM 6 and RSTAB 9 in VDI and Citrix Environments: Guide to Optimal Configuration

Introduction

The world of work is subject to constant change, in which flexibility and location-independent working are becoming increasingly important. Virtual Desktop Infrastructure (VDI) and Citrix solutions are an attractive way for companies to give employees access to their familiar working environment from anywhere in the world. These technologies centralize computing power and data management in a secure data center, which leads to simpler administration and improved data security.

However, VDI environments originally designed for standard office applications are not easily suited to the demanding requirements of 3D modeling and analysis programs, such as RFEM 6 and RSTAB 9. The intensive use of graphics power, the high number of read and write accesses, and the high demand for computing power for calculations can lead to significant performance issues in virtualized environments.

It is important to understand that the use of a VDI is always associated with certain performance losses compared to an installation on a physical high-end workstation. If the primary goal is to provide a system with absolute maximum computing speed for extensive simulations, the classic physical computer is the better choice.

This article therefore focuses on the technical levers that can be used to keep these system-related performance losses as low as possible in order to allow for efficient work in virtualized environments.

This guide is aimed at IT managers, system integration specialists, and experienced users who are faced with the challenge of implementing Dlubal software in the VDI or Citrix infrastructures. The aim is to analyze the specific technical challenges, explain the causes of performance issues, and present concrete, tried-and-tested solutions. A particular focus is on the strategic balance between cost, performance, and scalability in order to create a smooth and productive working environment for engineering professionals.

Basics: Virtual Desktop Infrastructure (VDI) and Citrix Technologies

What is Virtual Desktop Infrastructure (VDI)?

Virtual Desktop Infrastructure, or VDI for short, is a technology that provides a virtualized desktop environment. In contrast to conventional physical desktops, where the operating system and applications are installed locally on the user's computer, with VDI the desktop environments are hosted and run on central servers in a data center or in the cloud. The end user only accesses this virtual instance via an internet-enabled end device (that is, thin client, laptop, or tablet). The end device only serves as a “window” to the virtual environment, while all computing power and data processing takes place on the server side.

This model provides a number of decisive advantages for companies. Firstly, IT administration is considerably simplified by centralization. Instead of having to apply patches and software updates to thousands of individual devices, IT departments can carry out these processes centrally and synchronously for all virtual desktops. This not only saves time and resources, but also ensures that all employees can work with identical, consistent and up-to-date software versions.

Another key benefit is the increased security. As sensitive company data is not stored on the local devices of employees, but remains in the secure data center, the risk of data loss in the event of device loss or theft is significantly reduced. This also makes it easier to adhere to strict compliance regulations, such as those set out in the European General Data Protection Regulation (GDPR).

The following table lists the market-leading providers and platforms that offer solutions for virtual desktop environments:

Citrix: Leading VDI Solution

Citrix is a leading provider in the field of virtualization and remote access technologies. The Citrix Virtual Apps and Desktops solution is a core part of the VDI ecosystem and is used to provide virtual applications and desktops in a secure and scalable environment. The basic functionality of Citrix is to stream the rendered graphical user interface and user interactions to the end device via a special protocol, such as HDX.

Citrix systems provide the option for multiple users to share a VM. This saves resources, especially main memory. They allow for centralized management, provide a consistent and secure user experience, and use resources efficiently, making them relevant for demanding IT environments.

However, the centralization of computing power and data that characterizes VDI and Citrix is also the root of the challenges faced by users of demanding high-performance applications. The separation of computer (server) and screen (end device) through a network inevitably leads to latency, which can be particularly disruptive in interactive graphics and computing applications. Furthermore, the bundling of all data on a central storage system (instead of on individual SSDs) leads to potential shortages in the range of input/output operations (I/O), known as the “VDI IOPS problem”. This conflict between the advantages of centralization and the resulting performance issues forms the central conflict, which is discussed in detail below.

Basics of Data Management in RFEM 6 and RSTAB 9

The successful operation of Dlubal software in a virtualized environment requires a basic understanding of internal data management. RFEM 6 and RSTAB 9 follow a specific, I/O-intensive pattern that has a governing influence on the choice of the right VDI configuration.

RFEM 6 and RSTAB 9 Files and Importance of Working Directory

RFEM 6 and RSTAB 9 files are essentially nothing more than compressed ZIP archives. This structure is crucial for understanding the subsequent performance problems. When a model file is opened, the content is not simply loaded into the working memory (which would be far too small for this in most cases), but unpacked into a temporary working folder. By default, this folder is located in the user profile under the path C:\Users\\AppData\Local\Dlubal, but can be moved to a different location using the program options.

The unpacking process results in a single model file, which may be relatively small as a compressed ZIP archive, being split into a large number of smaller files. This generates a large number of initial write accesses to the working directory.

Description of I/O Operations

The special features in the file management of RFEM 6 and RSTAB 9 are not limited to the one-time unpacking. The entire workflow within the graphical user interface (GUI) and the solver is characterized by a continuous and intensive data exchange with this temporary working folder.

• GUI interaction: Dialog boxes load data directly from the working folder each time they are opened. After closing the dialog box, the changes made are written back to the folder. This process is constantly repeated and leads to a constant stream of small read and write accesses. Furthermore, working in the graphics window, switching visibilities, and numerous other actions also generate continuous I/O load.
• The calculation process: The calculation in RFEM 6 is by far the most I/O-intensive process. When the calculation is started, the solver line analyzes the calculation tasks and then starts several solver processes. Ideally, the number of these processes corresponds to the number of available CPU threads in order to make optimum use of the computing power. All of these solver processes simultaneously read data from the working folder and write back large amounts of data. This simultaneous access of several processes to the same memory location can lead to a so-called “I/O storm”, which extremely overloads the memory system.

The type of data processing in RFEM 6 is characterized by a high number of small files and intensive, simultaneous read and write accesses. Based on these characteristics, the program reacts particularly sensitively to the typical I/O bottlenecks in VDI environments. This specific problem profile requires targeted optimization strategies, which are discussed in detail in the following section.

Challenges Related to OpenGL

RFEM 6 and RSTAB 9 are dependent on OpenGL 4.2 for the graphical display. In many VDI and Citrix environments, this represents a considerable technical hurdle. Conventional virtualization solutions use simple software renderers by default, which process the graphics calculations purely via the CPU.

These standard software renderers often do not have the necessary OpenGL functions. If RFEM 6 or RSTAB 9 finds such a simple software renderer at program start, a corresponding warning is issued.

It is important to note that ignoring this warning may cause RFEM 6 or RSTAB 9 to crash immediately. Even if a graphical display is achieved, the processing of complex 3D scenes on the CPU leads to noticeable delays in the user interface. Users then experience a slower response from the model when rotating or zooming, which makes it much more difficult to work efficiently. The challenge is therefore to create an environment that not only provides the required OpenGL version, but also processes it efficiently.

Solutions for Performance Optimization

To ensure stable and efficient operation, it is necessary to address both the I/O bottlenecks and the graphics requirements.

Solutions for I/O Issue

The performance of file management can be significantly improved by making specific configuration changes and choosing the right hardware infrastructure.

• Defining antivirus exceptions: One of the most efficient ways to speed up is to disable real-time monitoring of security software for critical paths. Since RFEM 6 and RSTAB 9 process thousands of small files in milliseconds during the calculation, each scan slows down the system. The following folders should be excluded from monitoring:
  • The program folder (default: C:\Program Files\Dlubal).
  • The temporary working folder (default: C:\Users\\AppData\Local\Dlubal).
• Optimizing storage configuration: The working folder should ideally be located on a local, quickly connected storage system on the host server. If the use of a central storage system (NAS/SAN) is unavoidable, it is necessary that the network for connecting the storage has a high bandwidth and, above all, is designed with extremely low latency. High latency between the VM and the storage system can slow down even the fastest hardware. If necessary, the path for the temporary working folder can be changed. The path is stored in the registry in the path Computer\HKEY_CURRENT_USER\Software\Dlubal\RFEM6 in the key WorkingDirectoryPath.

Solutions for OpenGL Issue

For a smooth 3D display, the use of hardware acceleration through GPU virtualization is the recommended solution.

• GPU Passthrough: Here, a physical graphics card is assigned exclusively to a VM. This provides the highest performance, but is cost-intensive and less scalable. This solution is only provided in exceptional cases.
• Virtual GPU (vGPU): A physical graphics card is split into several virtual units and made available to several users simultaneously.A beginner's guide to workstation virtualization - DEVELOP3D This is the most efficient solution for most engineering workstations, as it provides a good balance between performance and user density.
• MESA software renderer as fallback: If GPU virtualization is not possible, the MESA renderer can be activated. This emulates the OpenGL functions on the CPU. Although this prevents crashes, the performance is considerably worse compared to hardware acceleration. The MESA renderer is activated via the script Enable Software Renderer.cmd in the program folder.

Conclusion and Comprehensive Recommendations

The successful operation of RFEM 6 and RSTAB 9 in VDI and Citrix environments requires strategic planning. RFEM 6 and RSTAB 9 place considerable demands on the performance of the hardware in terms of both interactive input and calculation. Especially for editing large models, a VDI solution is not always the optimal choice from a performance perspective.

In summary, the following conclusions can be drawn:

• The special type of data management makes the program susceptible to I/O bottlenecks.
• Deactivating antivirus real-time monitoring for specific folders may lead to a massive increase in performance.
• The separation of computing power and graphics requires efficient GPU virtualization.

Implementation is a problem that requires careful planning and testing. Expensive hardware can lose its effectiveness if configuration errors slow down performance.

Planning Checklist:

• Hardware sizing: Provide host servers with sufficient resources (CPU, RAM, GPU, I/O).
• Network configuration: Minimize latency between the end device and the server.
• Software tuning: Configure antivirus exceptions and optimize VMs.
• Monitoring: Regularly monitor workload to detect bottlenecks at an early stage.

With the right planning, the benefits of VDI can also be utilized for RFEM 6 and RSTAB 9.

The following table helps to localize and resolve problems that may arise in connection with VDI.