Building the world’s first network-aware Quantum Compiler that unlocks quantum computing at scale

The promise of quantum computing is immense, holding the potential to revolutionize industries from medicine to finance. However, a significant hurdle remains: building quantum computers large enough and reliable enough to tackle real-world problems. 

Today's quantum processing units (QPUs) are limited to hundreds of qubits, a far cry from the millions needed for practical, fault-tolerant quantum computation.

At Cisco Quantum Labs, we believe the most practical path to scalable quantum computing lies in distributed quantum computing, housed within a quantum data center (QDC).

This architecture networks multiple quantum processors together, allowing for a distributed system that can grow to meet the demands of future applications. It's not just about raw power; QDCs offer economic and operational benefits by centralizing quantum resources in a controlled environment.

The challenge: Bridging hardware and software in a distributed quantum world

Building a QDC isn't just about connecting QPUs. It requires a fundamental shift in how we design, manage, and operate quantum systems: multiple QPUs, multiple vendors, multiple modalities, multiple generations. This is where hardware-software co-design becomes critical.

A key component of this vision is a distributed quantum computing compiler that explicitly understands and leverages network connectivity and inter-processor communication.

Cisco Quantum Labs and Outshift by Cisco are pleased to introduce an industry-first network-aware Quantum Compiler prototype, purpose-built for quantum data centers. 

This is a new, foundational technology designed to accelerate the realization of scalable, fault-tolerant, and efficient quantum data centers. 

And the Quantum Compiler is now available as a free download for researchers and those looking to plan their future quantum infrastructure with accuracy, heterogeneity, and scale in mind.

Achieve scalable computation with the network-aware Quantum Compiler

Our Quantum Compiler prototype introduces several unique capabilities, making it a platform for the future of distributed quantum computing. The compiler’s key features address the major problems in scaling out a QDC, from splitting algorithms to orchestrating complex operations across devices. 

1. An industry-first distributed quantum error correction:

   1. The problem: Quantum computers are inherently noisy. Error correction is vital for fault-tolerant operation, but how do you implement it effectively across a distributed system?
   2. Our solution: Our compiler can encode circuits using error-correction codes (e.g., bivariate bicycle or surface code) and integrate them into the compilation and scheduling pipeline. This end-to-end flow allows us to quantify the impact of error correction on network resource utilization, including entanglement demand, communication cost, and execution latency.

2. Network-aware circuit partitioning:

   1. The problem: How do you split a quantum algorithm across multiple QPUs efficiently, minimizing the costly communication between them?
   2. Our solution: Our compiler intelligently partitions quantum circuits by viewing them as interaction graphs. It uses advanced techniques like windowed partitioning, modified Kernighan-Lin or multilevel partitioners, and selective state teleportation. Crucially, it only incurs inter-window qubit remapping when the benefits outweigh the movement overhead. This approach dramatically reduces entanglement consumption—by up to an order of magnitude compared to static partitioning.

3. Qubit mapping:

   1. The problem: Once a circuit is partitioned, how do you assign logical qubits to physical qubits across different QPUs, considering the network connectivity and performance?
   2. Our solution: The compiler maps qubits onto hardware locations that preferentially use the most reliable interconnects. It considers interaction graphs annotated with gate frequencies and network characteristics like link fidelity, entanglement generation rates, and bandwidth. Using an Integer Linear Programming (ILP) formulation, it delivers a globally optimized, network-aware assignment that preserves circuit fidelity.

4. Advanced scheduling:

   1. The problem: Orchestrating quantum operations and entanglement generation across a distributed network is complex, especially under hardware constraints.
   2. Our solution: The compiler's scheduler determines when to generate entanglement and execute nonlocal gates efficiently. It accounts for network topology, available communication qubits, and probabilistic entanglement generation. It supports both static scheduling (for easier analysis) and dynamic scheduling (for improved resource utilization and responsiveness).

5. Multi-tenancy for QDCs:

   1. The problem: In a shared QDC environment, how do you efficiently manage and schedule multiple quantum jobs competing for limited QPUs and network bandwidth?
   2. Our solution: We've developed a constraint-aware resource allocation algorithm and job reordering heuristics based on factors like qubit count and circuit depth. Simulation results show that strategic job ordering and QPU placement significantly reduce total execution cost and improve system utilization, paving the way for practical QDC operations.

Our Quantum Compiler is developed on a unified Quantum Networking software stack leveraging common protocols and algorithms for entanglement distribution, swapping, teleportation, and quantum measurement.

A co-design platform for the future of quantum infrastructure

Beyond its compilation capabilities, Cisco's Quantum Compiler also serves as a valuable co-design platform, as, in addition to its plug-in design, the unified network software stack it’s developed on also allows for emulated and simulated devices to participate in your experimentation and innovation needs:

• For algorithm developers: The compiler's toolkit enables detailed performance modeling and optimization. Developers can assess trade-offs between qubit allocation, network latency, error rates, and resource scheduling, adapting their algorithms for heterogeneous, interconnected quantum processors.
   
• For quantum data center designers: The tools provide capabilities for designing and validating QDC architectures. Designers can model and evaluate infrastructure based on target QPU technologies, network topologies, and intended applications, optimizing for maximum scalability, reliability, and efficiency.

While the compiler comes equipped with Cisco’s own suite of algorithms, it is designed to be extensible. Researchers and developers can integrate their own algorithms to add new functionalities within the controller, and seamlessly test them through the compiler’s pipeline.

By holistically coupling software, networking, and hardware layers, Cisco's Quantum Compiler prototype is a critical enabler for the next generation of Distributed quantum data centers, bringing us closer to the era of scale-out practical, fault-tolerant quantum computing.

Try the Quantum Compiler now                                      

Now available for download from Cisco Quantum Labs and Outshift by Cisco, the Quantum Compiler can help you determine how many quantum computing nodes you'll need and what types of compute technologies work is the best architecture of distributed quantum computing for your quantum algorithm.

We think the possibilities for enterprise application are endless:

• Pharma companies doing drug discovery with quantum algorithms too complex for single machines
• Financial firms running quantum optimization algorithms that need more computational power and to right-size their infrastructure
• Research institutions innovating new quantum algorithms and compute types
• Hyperscalers or high-performance computing providers to plan infrastructure for their quantum data center with their choice of quantum computing technology

Design the blueprint for your quantum infrastructure needs without the guesswork. 

Access our free Quantum Compiler prototype here: You’ll have 30 days to try it out, and you may be contacted by our team to learn about your experience.