Microsoft has made a major rewrite of their Azure Quantum Development Kit (QDK) to make it 100X faster, 100X smaller, and able to run in a browser. The new version has just been released as the Azure Quantum Development Kit 1.0. The key change driving the improvements is that the internal code base was converted to the Rust programming language known for its higher efficiency in runtime and memory usage. It also includes a new of new or improved features such as full IntelliSense support which is a code completion aid that makes coding more convenient, built-in quantum simulation and debugging, integrating many of the capabilities of the Resource Estimator into the QDK, an providing refine Python packages and Jupyter notebook support. There are a few incompatibilities in this new QDK, which Microsoft is calling their “Modern QDK” versus the earlier one, which Microsoft now calls their “Classic QDK”. So users who had projects in the previous QDK will need to make some minor changes to convert to the new QDK.

Additional information about Microsoft’s new Azure Quantum Development Kit 1.0 release can be found in a blog post announcing it here, an article that provides detail on what’s new in the QDK is here, a page with installation instructions here, and another blog article that goes into the features and internal design in greater depth here.

The biggest changes in the Azure Quantum Resource Estimator is that Microsoft has now made it open sourced and its functionality has been integrated with the Azure QDK mentioned above. This should make it easier and faster for users of the new Azure QDK to try out. The Resource Estimator is specifically oriented for exploring how well quantum algorithms will run on various fault tolerant quantum machines. A user will need to input key parameters such as qubit parameters, quantum error correction schema, T-state distillation parameters, etc. and the program will make an estimate of the number of qubits required and the overall runtime. If a user doesn’t want to enter in their own parameters, Microsoft has available six pre-defined sets of parameters that cover superconducting, ion trap, and Majorana based qubit technology. These parameters are, of course, critical for estimating performance, but its not clear how easy it will be to estimate the parameters accurately. New quantum processors with different characteristics are constantly being released on a regular basis and there also could be high variations from vendor-to-vendor even when both vendors are using the same modality. But the estimator should still be useful for an end user to get a rough feel for what type of machine will be needed to support that algorithm they are developing.

A blog post from Microsoft announcing the updated Resource Estimate can be viewed here and another post that provides users with additional details on how to customize the target parameters of the resource estimator is available here.

January 20, 2024