Skip to content
Logo

pytket-custatevec

GPU-accelerated statevector and shot-based simulation for pytket.

Get Started View Examples
Build PyPI License Downloads Slack StackExchange

pytket-custatevec acts as an open-source bridge between Quantinuum's pytket compiler and NVIDIA's cuQuantum SDK, enabling massive speedups for circuit simulations.

Why use this backend?

  • High Performance --- Leverage NVIDIA GPUs to enable rapid simulation of quantum circuits using parallel architecture.

  • Seamless Integration --- Works as a standard pytket Backend. Just switch your backend import, and your existing code runs on the GPU immediately.

  • Optimized Memory --- Utilizes cuStateVec's advanced memory management to handle large statevectors efficiently on GPU VRAM.

Architecture

This library sits directly on top of the NVIDIA cuQuantum stack.

flowchart LR %% Node Definitions User([User Code]) Backend[CuStateVecBackend] SDK[NVIDIA cuStateVec] GPU[NVIDIA GPU] %% Connections User == pytket Circuit ==> Backend Backend == cuQuantum Python ==> SDK SDK -.-> |CUDA| GPU %% Assign Classes (CSS will handle the colors) class User user; class Backend pytket; class SDK nvidia; class GPU hardware;

Bugs and Support

This is a community-driven project. Please file bugs and feature requests on the GitHub Issue Tracker.