PsiQuantum is launching Construct, our software suite for designing, developing, and optimizing fault tolerant quantum algorithms. Construct represents years of development by our quantum applications team, who created these tools for research into new algorithms for PsiQuantum’s first generations of fault tolerant quantum computers. We are excited to share these tools with the outside world to support their further research into fault tolerant quantum algorithms and applications.

Why build a new quantum development platform?

PsiQuantum was founded with the belief that the true potential of quantum computing would only be realized with utility-scale quantum computers: fault tolerant systems in the order of a million physical qubits. However, building a utility-scale quantum computer is only half the picture. To achieve utility, we also need fault tolerant quantum algorithms designed to take advantage of this new class of hardware.

When we began our journey to develop fault tolerant algorithms, we found many quantum development tools designed to tackle small research problems for NISQ hardware with dozens of qubits and hundreds of operations. These tools focused on supporting low level qubit manipulation, noise modeling, and error mitigation. However, when used to develop algorithms at the scale of fault tolerant machines, these tools were not up to challenge. Fault tolerant algorithms, even for the early generation of machines, must support hundreds of logical qubits and billions of operations. Error mitigation is largely not a concern, and high-level programming abstractions and resource optimization become critical. To support our own algorithmic development, we were forced to develop our own software toolkit.

What is Construct?

Last year we were excited to share one piece of our quantum development toolkit with the research community: Bartiq and QREF. Now with the launch of Construct our partners can access our core algorithm development tools. Construct is our externally accessible software platform for designing, developing, and optimizing fault tolerant quantum algorithms, built originally to support PsiQuantum’s algorithmic research teams.

Circuit Designer

Circuit Designer is an intuitive visual editor for quantum circuits built directly into Construct. It lets developers quickly assemble and document circuits on a flexible canvas — ideal for prototyping, sharing, and refining your ideas.

We’ve reimagined how to build circuit diagrams from the ground up to support how quantum researchers develop fault tolerant algorithms:

Circuit Designer is Fast: Seamlessly add hundreds of gates to your circuits without performance issues or slowdown.

Circuit Designer is Friendly: Quickly sketch circuits by simply dragging and dropping gates, registers, and other components onto your canvas.

Circuit Designer is Flexible: Easily organize gates into collapsible routines for better clarity and modularity.

Circuit Designer is Familiar: Clearly label circuit elements using familiar LaTeX notation for consistent and professional documentation.

Workbench within our Quantum Development Environment (QDE)

Workbench is our Python library for developing fault tolerant quantum algorithms. Fault tolerant algorithms, representing billions of operations across hundreds of logical qubits, require working at a much larger scale than traditional NISQ algorithms of hundreds of operations on dozens of physical qubits. Workbench has many features designed from the ground up to support the scale of fault tolerant algorithms.

One of these features is the ability to build modular subroutines that can be easily reused across algorithms. This lets FTQC researchers repurpose existing code in new contexts, avoiding redundant work.

For instance, the current state-of-the-art block encoding scheme in quantum chemistry, DF-THC, extends earlier double factorization (DF) and tensor hypercontraction (THC) methods with a new Hamiltonian decomposition. Since we already had Workbench implementations of DF and THC, we were able to implement DF-THC soon after the paper appeared. By reusing our gate-based implementations of Givens rotations (via phase gradient addition), coherent alias sampling state preparation, and quantum phase estimation, and combining them with “black box” implementations of DF-THC data loaders, we implemented the algorithm and validated Quantum Resource Estimates (QRE) against the paper’s results within weeks. With this implementation, we could automatically produce estimates for the active volume of the algorithm in a photonic architecture, as well as explore the design space in trading off logical qubits against logical operations.

Our Quantum Development Environment (QDE) is a hosted and managed development environment that allows you to quickly get started developing quantum algorithms with Workbench and other helpful Python libraries for quantum computing research. With our QDE you can begin coding from your web browser without installing and managing any software on your own computer. We support access to our QDE through both VSCode and Jupyter notebook interfaces.

Resource Analyzer

Resource analyzer is a visual tool to breakdown a quantum algorithm’s resource costs to uncover opportunities for optimization. Through different visualizations, it shows the sub-routine organization of your algorithms and which sub-routines are consuming the most resources. It allows quick visual inspection to uncover resource hotspots for different resource measures, guiding developers towards opportunities for further optimization.

How can partners access Construct?

Today we are announcing our “Early Access Program” (EAP) for Construct. As part of EAP, our close partners can begin exploring Construct for their own fault tolerant algorithmic research. By engaging with Construct in its earliest form, our partners can explore the first public version of our tools and provide feedback to shape their future direction.

We are also excited to support community access to Construct during EAP. If you are interested in exploring Construct, please express your interest here. As we are still in the early stages of sharing Construct, we will only be supporting a select number of community users at this time. Construct is designed for quantum computing experts and as such we are limiting access to experts in the research community who can take full advantage of our platform's capabilities.

What’s next?

We are excited to share our quantum algorithmic developer platform with our partners and members in the research community. We hope they will find our tools as useful as our own teams have. But this is just a start for us. We have several new tools coming soon to Construct to make fault tolerant quantum programming more accessible to scientists who are eager to explore this new computing capability and look forward to expanding Construct access to the broad research community after our Early Access Program concludes. Together we can work towards developing useful quantum algorithms and applications for the first generation of utility scale quantum hardware.

Learn more

Introducing Construct - Inside PsiQuantum's Software Suite