💚 CUDA-Q

CUDA-Q is NVIDIA's open-source platform for quantum-GPU computing. Write a quantum program once in Python or C++ using @cudaq.kernel, and run it across NVIDIA's GPU-accelerated simulators or real QPUs without rewriting your code.
What you get at a hackathon:
Single-source hybrid programming. Express quantum and classical logic together in one language you already know, with tight integration to classical compute.
GPU-accelerated simulation through cuQuantum. State vector, tensor network, density matrix, and more, scaled across multiple GPUs and nodes so you can simulate larger systems and run faster than CPU-only tools.
Hardware portability. The same kernel targets simulators or physical QPUs from multiple hardware partners. Prototype on a GPU, point at a real device when you're ready.
Accelerated libraries. cuQuantum for simulation, CUDA-Q QEC for quantum error correction research including the first GPU-accelerated decoders, and CUDA-Q Solvers for application building blocks.
An MLIR-based compiler that handles optimization and lowering, so you focus on the algorithm.
Pitch line: write it once, simulate it fast on GPUs, and run it on real quantum hardware, all from Python or C++.

Closed bounties:

• $100 | SABRE qubit-mapping pass: topology-aware initial placement to reduce SWAPs on irregular devices
  closed by: border-b
• $100 | Propagator computation for dynamics
  closed by: friedsam
• $100 | Add common quantum embeddings
  closed by: simsaidan, ssmswapnil
• $100 | Mathematical functions support
• $100 | [custom op] Support unitary synthesis for 3+ qubit operations
  closed by: thedaemon-wizard