ANSYS CFD scalability on Microsoft Azure is outstanding

“ANSYS and Microsoft Azure have been working closely on a Proof of Concept (POC) with a large customer to run ANSYS CFD workload on Azure,” said Ray Milhem, vice president of enterprise solutions at ANSYS. “The POC proved very successful and the data showed excellent scalability running ANSYS CFD up to 1024 cores.”


Azure for Engineers

Do you rather spend your time installing and configuring hard- and software or trying to get some work done? Many companies are currently switching off their servers and closing down the air-conditioned server rooms. The new IT infrastructure runs in the cloud and is cheaper, more efficient, and better for the environment.


A good example of successful use of Cloud computing was a Master Thesis at the HSR University of Applied Sciences in collaboration with the Nanyang Technical University (NTU) in Singapore. The task was to model the wing motion of a dragonfly using fluid-structure interaction in ANSYS Workbench. Since NTU did not have access to sufficient computational power, a Virtual Workstation with ANSYS 16.1 was set up in Microsoft Azure. The computer could be accessed both from Switzerland and Singapore and the 900 GB of data did not have to be stored in either place.

If you are interested in using Microsoft Azure for running simulations in Cloud, please get in touch with us. You can also visit us at the Swiss Symposium for Virtual Product Development at the HSR in Rapperswil on April 27, 2016.

Best regards,
Henrik Nordborg

Heating Load Prediction in AzureML

This short movie shows an example of prediction possibilities of Microsoft AzureML platform. The input data set is composed of ca. 700 buildings each representet by a feature set. The goal of the prediction is to predict so called heating load of the building.

The prediction is performed with three algorithms: decision forrest regression, linear regresion and linear regression with permutated features. The ‘score’ module presents the prediction quality by comparing original and predicted values.

How to automatize VM deployment


If you have customers asking for the VMs with the same software suite, sooner or later the full automatization of the deployment is required. Powershell Microsoft Azure API is probably the simplest and fastest way to programatically deliver the VM with preinstalled software.

Let’s say you have a customer that wishes a VM with Software X, access to storage service and fixed IP. You have to start with an empty subscription.

Here we provide a short instruction set how this automatization can be realized.

  1. Deploy a VM and install necessary software. This step can be performed once in the Azure Portal.
  2. Generalize the image of the VM. Write down the full URI of the image. It will be stored in the Azure storage service as a blob.
  3. Create the target virtual network.
  4. Execute the PowerShell script that deploys a new VM and uses the generalized image as an OS Disk.


Note, that the script creates also credentials for you, is deployed in the existing virtual network, and also assigns the VM to the network interface with a public fixed IP.

How to run OpenFOAM in Microsoft Azure


This tutorial describes how to deploy a VM with Ubuntu in Microsoft Azure and run simulations in OpenFOAM. It is assumed that these steps are performed on a local Windows machine.

1. Create the Virtual Machine with Linux

  1. Register the Azure Subscription.
    1. You should receive the PromoCode from your teacher in a separate email. Redeem the code at
    2. For the registration you should use your LiveID Account, e.g., If you don’t have such an account, please register.
  2. Login to Mirosoft Azure Portal at
  3. Create the virtual machine with Ubuntu Server 14.04 LTS.New->Compute->Ubuntu Server 14.04 LTSCapture
  4. Follow this instruction and use default settings. If you don’t have a SSH public key use user/password.
    Use Resource Manager and North Europe region.
  5. Pick the D13 size of the VM with 8 cores and 56 GB RAM. For initial testing you can D12 with 4 cores and 28 GB RAM.
  6. Note the IP address of your VM.

2. Connect to the VM and Configure Visualization

Now, you can access your VM with so called SSH Protocol. In Linux and OSX ssh clients are available by default. If you use Windows you have to install a special ssh client such as Putty.

Note: when connecting to a remote Linux VM over SSH the display is not available. One possible solution is to install a local X-Windows Server and enable X Tunelling in SSH Client. This way the X-Window instructions are tunelled over SSH to the on-premises X-Server. To enable this configuration follow these instructions and turn on the X11 Tunelling in Putty. In particular:

  1. Install and run Xming Server from Sourceforge.
  2. Install Putty (if not previously installed).
  3. Enable X11 Tunelling in Putty.
  4. Run Putty and connect to the VM using its IP address you previously noted.
  5. Run xclock or xemacs to test if the remote visualisation works.

Now you can run Paraview remotely. In addition you can 1) fix your name in Putty to ease to log in (Putty->Connection->Data), 2) enable compression (Putty->SSH->Enable Compression).

3. Install OpenFOAM

  1. Follow these instructions and install OpenFOAM 3.0.1 and Paraview using apt package.step2.PNG
  2. Configure OpenFOAM and make sure that OpenFOAM works
    Important: paraview and parafoam will not work at this step.step3

Note: obviously gedit editor is not available. Follow the instructions below to install another editor emacsVIM is always available (just type in vim in the terminal).

4. Install Editor

You can always work with vim. If you prefer another editor, you can install it using apt-get. This is an example for Emacs:

sudo apt-get install emacs24

emacs &

5. German Localisation

If you are unconfortable with US keyboard, you can switch it to German by typing:

sudo dpkgreconfigure keyboardconfiguration

Select Generic 105 keys, Origin: German, Keyboard layout: German. All other options can remain default. Then reboot the machine in the Azure Portal.

6. Configure Storage

The VM comes with the local hard drive that has a limited capacity. One of Microsoft Azure Services provides persistent and redundant storage service. The storage service supports SMB Protocol and can be mapped to a drive. The instructions how to do that can be found here. In particular you have to:

  1. Create a File Storage Service using the Azure Portal.
  2. Create a container.
  3. Follow these instructions to map the file share as local drive.

Now using a client on-premises system such as CloudBerry Explorer you can upload the data to the Azure File Storage. Once the data are uploaded they become immediately available in the cloud.


7. Have Fun