1. QTCAD

This section describes how to install QTCAD and its dependencies.

1.1. Software requirements

QTCAD consists of three main Python libraries, and thus requires a properly working Python installation. The recommended Python distribution for QTCAD is Miniconda, which may be installed using the instructions on the official conda website. This will include both Python and conda, the preferred package management tool for QTCAD installation.

QTCAD also uses mesh files as input. These mesh files can be produced with Gmsh. Following the conda QTCAD installation instructions below ensures that Gmsh is properly installed in the QTCAD virtual environement. In any situation where this is not possible, Gmsh may be installed by following the instructions on the official Gmsh website or in the next section. The website also contains Gmsh tutorials. Video tutorials on how to use Gmsh can also easily be found on YouTube.

Similarly, 3D mesh and geometry files may be generated from 2D layouts using the Python API coded in devicegen. Again, the conda installation procedure explained below will install this package automatically. Manual installation of devicegen is explained in the devicegen installation page.

Remark that while QTCAD is a commercial proprietary software, all its dependencies specified above are available for free.

1.2. Hardware requirements

Precise hardware requirements are strongly dependent on the complexity of the problem being solved both in terms of size of the mesh and solvers being employed. Therefore, it is quite non-trivial to provide universal hardware requirements for QTCAD.

Instead, below, we give two examples of systems on which QTCAD simulations may be run successfully. The first system is a laptop which may be used to run basic simulations for meshes with up to approximately 200,000 nodes that define simple devices with one or two quantum dots. This system uses a processor that dates back from 2015 and should be interpreted as a minimal setup. The second system is a more advanced desktop workstation which we routinely use for complex simulations using meshes with up to approximately 2,000,000 nodes.

These configurations should be understood as typical system examples, and not as hard requirements. We have indeed successfully run QTCAD on a wide variety of systems (Windows, Mac, or Linux laptops, desktop workstations, cloud computing services, HPC clusters, etc.). Since computation times depend non-trivially on many factors such as the hardware being used, number of cores employed, solver, mesh dimension, convergence properties of the problem being studied, tolerance of the solvers, or basis-set dimension (in the case of the many-body solver), they are not given here. Instead, we encourage users to experiment with our tutorials and practical applications on their machine. Nevertheless, for simulations of typical few-electron single or double quantum-dot devices, a laptop or a desktop computer should suffice, and HPC clusters should not be necessary.

1.2.1. System 1 (laptop)

CPU	6th Gen Intel^® Core™ i5-6300U @ 2.40GHz (4 threads)
GPU	No discrete GPU, only integrated graphics (GPU is used only for plotting in ParaView and mesh rendering in Gmsh)
RAM	8 GB
HDD	256 GB (SSD)
OS	Windows 10

1.2.2. System 2 (desktop)

CPU	13th Gen Intel^® Core™ i9-13900K @ 3.00 GHz (32 threads)
GPU	None
RAM	128 GB
HDD	1 TB (SSD)
OS	Linux (Ubuntu distribution)

1.2.3. Operating system

We officially support the following operating systems:

Windows
Linux
macOS

1.3. Downloading QTCAD

To download QTCAD, please visit the Nanoacademic client portal at

https://portal.nanoacademic.com/

If you already have a Nanoacademic account, please log in using your password.

Otherwise, please create your account, and then log in.

If this was not already done, please proceed to fill in the required user profile information (name, email address, institution, phone number, etc.).

Once you have a valid profile, please click on the “Products” section and select QTCAD. You should then see the following page:

Getting a free trial version — Fig. 1.3.1 Getting a free QTCAD trial version

Click on “Start your free trial”.

You will then see the following page which will allow you to download your license.

Fig. 1.3.2 Downloading a free trial license of QTCAD

This will produce a file called license.lic which we discuss further in the License Activation page.

The next step is to click the “Download QTCAD” button. This will bring you to a page containing all available QTCAD versions. Please download the latest one at the top of the page.

1.4. Package installation with conda (recommended)

1.4.1. Setting up conda

Assuming that Miniconda (or Anaconda) is installed on your machine as recommended in the Software requirements section, it is good practice to make sure that you are using the latest version of conda. This can be done by updating your base conda environment.

conda update -n base conda

In addition, though this is optional, we strongly recommend installing Libmamba in your base environment. Libmamba is a faster dependency solver for conda which can significantly speed up QTCAD installation. Libmamba can be installed and set as the default dependency solver in your base environment as follows:

conda install -n base conda-libmamba-solver
conda config --set solver libmamba

1.4.2. QTCAD installation on Linux and macOS

Unzip the compressed file in the target installation directory (represented here by installation_path).
```
unzip qtcad.zip -d installation_path
```
Move to the qtcad directory inside the installation directory.
```
cd installation_path/qtcad
```
Create the virtual environment that will be used to run QTCAD.
```
conda env update -f qtcad-OS.yml
```
where OS stands for linux or mac, depending on your operating system.
Warning

If you are using a Mac with an Arm-based processors (e.g. Apple M1), please install an osx-64 version of the environment using Rosetta due to the lack of support of some packages for Arm-based processors.
softwareupdate --install-rosetta ** RESTART YOUR SHELL ** CONDA_SUBDIR=osx-64 conda env create -f qtcad-mac.yml
Activate the virtual environment.
```
conda activate qtcad
```
Install the package
```
pip install .
```

Once the package is installed, it can be used like any regular Python package, provided the qtcad virtual environment is activated (step 4, above).

1.4.3. QTCAD installation on Windows

To install QTCAD in Windows, it is recommended to use a PowerShell terminal.

Unzip the compressed directory in the target installation directory (represented here by the string installation_path).
```
Expand-Archive qtcad.zip installation_path
```
Move to the qtcad directory inside the installation directory.
```
cd installation_path/qtcad
```
Create the virtual environment that will be used to run QTCAD.
```
conda env update -f qtcad-win.yml
```
Activate the virtual environment.
```
conda activate qtcad
```
Install the package
```
pip install .
```

Once the package is installed, it can be used like any regular Python package, provided the qtcad virtual environment is activated (step 4, above).

1.5. Package installation with pyenv (not recommended)

It is strongly recommended to install QTCAD using conda (Section 1.4).

However, some high-performance computing (HPC) clusters may not allow a conda install for various reasons (technical or administrative).

In these situations, installation with pyenv may be attempted despite the above recommendation.

If you need to install QTCAD through pyenv or any approach different from the default conda installation method, please contact Nanoacademic to get the required support. For example, the requirements.txt file included with QTCAD for pyenv installations may have to be manually tailored to your workstation and operating system.

1.5.1. Linux and macOS

1.5.1.1. Installing pyenv

Before installing pyenv, it is useful to make sure that the recommended build environment is first put in place.

To install pyenv, an automatic installer is available on the official pyenv GitHub webpage.

If you prefer to manually clone the pyenv git repository, please follow these instructions, and then make sure to correctly set up your shell environment according to the procedure described after the section linked above.

After installing pyenv, a necessary next step to proceed with the QTCAD installation instructions that follow is to install the pyenv-virtualenv plugin of pyenv.

Note that if the automatic installer was used, this plugin may already have been installed. In this case, assuming that pyenv was installed in a standard directory, a directory called pyenv-virtualenv will be found in $HOME/.pyenv/plugins/.

1.5.1.2. Installing QTCAD dependencies

Install the Python 3.8 interpreter with
```
pyenv install 3.8.0
```
Create a virtual environment for QTCAD using the Python 3.8 interpreter with
```
pyenv virtualenv 3.8.0 qtcad
```
Activate the qtcad environment with
```
pyenv shell qtcad
```
The list of virtual environments can be shown with
```
pyenv virtualenvs
```
A typical output will look like
```
3.8.0/envs/qtcad (created from /home/user/.pyenv/versions/3.8.0)
* qtcad (created from /home/user/.pyenv/versions/3.8.0)
```
Here, the asterisk (*) indicates that the active environment is the qtcad environment.
Install the packages specified in the requirements.txt file using pip with
```
pip install -r requirements.txt
```

1.5.1.3. Installing QTCAD

Finally, QTCAD itself is installed with

pip install .

After installation, QTCAD may be launched simply by activating the qtcad virtual environment (step 3 in Section 1.5.1.2). It should then be possible to run Python scripts that call any of the QTCAD packages.

1.5.2. Windows

pyenv is not officially supported in Windows. A Windows version called pyenv-win exists, but was not tested by Nanoacademic. Please contact Nanoacademic if you need to install QTCAD on Windows with any approach other than the recommended conda approach.