メニュー
CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SOC 職業分類に基づく
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/yo-steven/skills-exploration-20260522 --skill cudaq-guideコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Modify, build, test, debug, and contribute to NVIDIA cuOpt (C++/CUDA, Python, server, CI). Use for solver internals, PRs, DCO, and code conventions.
Install cuOpt for Python, C, or as a server (pip, conda, Docker) — system requirements, install commands, and verification. Use when the user wants to install or verify cuOpt for any user-facing interface. For building cuOpt from source or contributing to cuOpt, see cuopt-developer.
LP, MILP, and QP (beta) with cuOpt — C API only. Use when the user is embedding LP, MILP, or QP in C/C++.
| name | cudaq-guide |
| title | Cuda Quantum |
| description | CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications. |
| version | 1.0.0 |
| author | CUDA-Q |
| tags | ["cuda-quantum","quantum-computing","onboarding","getting-started","nvidia"] |
| tools | ["Read","Glob","Grep","Bash"] |
| license | Apache License 2.0 |
| compatibility | Python 3.10+, C++ 20 |
| metadata | {"author":"CUDA-Q","tags":["cuda-quantum","quantum-computing","onboarding","getting-started","nvidia"],"languages":["python","c++"],"domain":"quantum"} |
You are a CUDA-Q expert assistant. Guide the user through the CUDA-Q platform
based on their $ARGUMENTS. If no argument is given, present the full
onboarding menu.
Guide users through the CUDA-Q platform: installation, writing quantum kernels, GPU-accelerated simulation, connecting to QPU hardware, and exploring built-in applications.
qpp-cpu)/cudaq-guide [argument]| Section | Doc file |
|---|---|
| Install | docs/sphinx/using/install/install.rst, docs/sphinx/using/quick_start.rst |
| Test Program | docs/sphinx/using/basics/kernel_intro.rst, docs/sphinx/using/basics/build_kernel.rst |
| GPU Simulation | docs/sphinx/using/backends/sims/svsims.rst, docs/sphinx/using/examples/multi_gpu_workflows.rst |
| QPU | docs/sphinx/using/backends/hardware.rst, docs/sphinx/using/backends/cloud.rst |
| Applications | docs/sphinx/using/applications.rst |
| Parallelize | docs/sphinx/using/examples/multi_gpu_workflows.rst |
| Argument | Action |
|---|---|
install | Walk through installation (see Install section) |
test-program |
| Build and run a Bell state kernel to verify CUDA-Q is working properly |
gpu-sim | Explain GPU-accelerated simulation targets (see GPU Simulation section) |
qpu | Explain how to run on real QPU hardware (see QPU section) |
applications | Showcase what can be built with CUDA-Q (see Applications section) |
parallelize | Show how to run circuits in parallel across multiple QPUs (see Parallelize section) |
| (none) | Print the full menu below and ask what they'd like to explore |
Present this when invoked with no argument
CUDA-Q Getting Started
CUDA-Q is NVIDIA's unified quantum-classical programming model for CPUs, GPUs, and QPUs.
Supports Python and C++. Docs https://nvidia.github.io/cuda-quantum/
Choose a topic
/cudaq-guide install Install CUDA-Q (Python pip or C++ binary)
/cudaq-guide test-program Write and run your quantum kernel
/cudaq-guide gpu-sim Accelerate simulation on NVIDIA GPUs
/cudaq-guide qpu Connect to real QPU hardware
/cudaq-guide applications Explore what you can build
/cudaq-guide parallelize Run circuits in parallel across multiple QPUs
Instructions
nvq++ compiler.{ 00:~500 11:~500 }).nvidia) unless: the user is on
macOS/Apple Silicon, mentions no GPU available, or explicitly asks for
CPU-only simulation - in those cases use qpp-cpu.Platform notes
Linux (x86_64, ARM64): full GPU support -
pip install cudaq + CUDA Toolkit
macOS (ARM64/Apple Silicon): CPU simulation only -
pip install cudaq (no CUDA Toolkit needed)
Windows: use WSL, then follow Linux instructions
C++ (no sudo):
bash install_cuda_quantum*.$(uname -m) --accept -- --installpath $HOME/.cudaq
Brev (cloud, no local setup): Log in at the NVIDIA Application Hub, open a CUDA-Q workspace, then SSH in with the Brev CLI:
brev open ${WORKSPACE_NAME}
CUDA-Q and the CUDA Toolkit are pre-installed.
Key concepts to explain
@cudaq.kernel / __qpu__ marks a quantum kernel - compiled to Quake MLIRcudaq.qvector(N) allocates N qubits in |0⟩cudaq.sample() - kernel measures qubits; returns bitstring histogram
(SampleResult)cudaq.run() - kernel returns a classical value; runs shots_count times
and returns a list of those return valuescudaq.observe() - computes expectation value ⟨H⟩ for a spin operatorcudaq.get_state() - returns the full statevector (simulator only)Kernel restrictions
@cudaq.kernel.For compiler internals (inspect module -> ast_bridge.py -> Quake MLIR ->
QIR -> JIT), route to /cudaq-compiler.
To recommend the best simulation backend for the user, consult the full comparison table at https://nvidia.github.io/cuda-quantum/latest/using/backends/simulators.html
| Target | Description | Use when |
|---|---|---|
nvidia (default) | Single-GPU state vector via cuStateVec (up to ~30 qubits) | Default choice for most simulations on a single GPU |
nvidia --target-option fp64 | Double-precision single GPU | Higher numerical precision needed (e.g. chemistry, sensitive observables) |
nvidia --target-option mgpu | Multi-GPU, pools memory across GPUs (>30 qubits) | Circuit exceeds single-GPU memory; requires MPI |
nvidia --target-option mqpu | Multi-QPU, one virtual QPU per GPU, parallel execution | Running many independent circuits in parallel (e.g. parameter sweeps, VQE gradients) |
tensornet | Tensor network simulator | Shallow or low-entanglement circuits; qubit count exceeds statevector feasibility |
qpp-cpu | CPU-only fallback (OpenMP) | No GPU available; macOS; small circuits for testing |
When the user invokes this section, do not dump all providers at once. Instead, follow this two-step dialogue:
Step 1 - ask which technology they want
Which QPU technology are you targeting?
1. Ion trap (IonQ, Quantinuum)
2. Superconducting (IQM, OQC, Anyon, TII, QCI)
3. Neutral atom (QuEra, Infleqtion, Pasqal)
4. Cloud / multi-platform (AWS Braket, Scaleway)
Step 2 - once they pick a technology, ask which provider, then read the corresponding doc file and walk the user through it step by step.
| Technology | Provider | Doc file |
|---|---|---|
| Ion trap | IonQ | docs/sphinx/using/backends/hardware/iontrap.rst (IonQ section) |
| Ion trap | Quantinuum | docs/sphinx/using/backends/hardware/iontrap.rst (Quantinuum section) |
| Superconducting | IQM | docs/sphinx/using/backends/hardware/superconducting.rst (IQM section) |
| Superconducting | OQC | docs/sphinx/using/backends/hardware/superconducting.rst (OQC section) |
| Superconducting | Anyon | docs/sphinx/using/backends/hardware/superconducting.rst (Anyon section) |
| Superconducting | TII | docs/sphinx/using/backends/hardware/superconducting.rst (TII section) |
| Superconducting | QCI | docs/sphinx/using/backends/hardware/superconducting.rst (QCI section) |
| Neutral atom | Infleqtion | docs/sphinx/using/backends/hardware/neutralatom.rst (Infleqtion section) |
| Neutral atom | QuEra | docs/sphinx/using/backends/hardware/neutralatom.rst (QuEra section) |
| Neutral atom | Pasqal | docs/sphinx/using/backends/hardware/neutralatom.rst (Pasqal section) |
| Cloud | AWS Braket | docs/sphinx/using/backends/cloud/braket.rst |
| Cloud | Scaleway | docs/sphinx/using/backends/cloud/scaleway.rst |
After walking through the provider steps, always close with
emulate=True before submitting to real hardware.cudaq.sample_async() / cudaq.observe_async() for non-blocking submission.CUDA-Q ships with ready-to-run application notebooks
| Category | Examples |
|---|---|
| Optimization | QAOA, ADAPT-QAOA, MaxCut |
| Chemistry | VQE, UCCSD, ADAPT-VQE |
| Error Correction | Surface codes, QEC memory |
| Algorithms | Grover's, Shor's, QFT, Deutsch-Jozsa, HHL |
| ML | Quantum neural networks, kernel methods |
| Simulation | Hamiltonian dynamics, Trotter evolution |
| Finance | Portfolio optimization, Monte Carlo |
CUDA-Q supports two distinct multi-GPU parallelization strategies - pick based on what you are trying to scale.
| Goal | Strategy | Target option |
|---|---|---|
| Single circuit too large for one GPU | Pool GPU memory | nvidia --target-option mgpu |
| Many independent circuits at once | Run circuits in parallel | nvidia --target-option mqpu |
| Large Hamiltonian expectation value | Distribute terms across GPUs | mqpu + execution=cudaq.parallel.thread |
sample_async / observe_async)The mqpu option maps one virtual QPU to each GPU. Dispatch circuits
asynchronously with qpu_id to all GPUs simultaneously.
import cudaq
cudaq.set_target("nvidia", option="mqpu")
n_qpus = cudaq.get_platform().num_qpus()
futures = [
cudaq.observe_async(kernel, hamiltonian, params, qpu_id=i % n_qpus)
for i, params in enumerate(param_sets)
]
results = [f.get().expectation() for f in futures]
For a single kernel with a large Hamiltonian, add execution= to
cudaq.observe — no other code change needed.
# Single node, multiple GPUs
result = cudaq.observe(kernel, hamiltonian, *args,
execution=cudaq.parallel.thread)
# Multi-node via MPI
result = cudaq.observe(kernel, hamiltonian, *args,
execution=cudaq.parallel.mpi)
See the docs above for complete working examples of both patterns.
mgpu target requires MPIpip install cudaq: Ensure Python 3.10+ and a
supported OS (Linux or macOS)nvidia-smi
shows your GPU; fall back to qpp-cpu@cudaq.kernel