Available Resources and Tools
JUNIQ Cloud provides a variety of options and features to enhance your interactive computing experience, including access to quantum computing resources. Here's an overview of what you can expect:
1. Systems Available
JUNIQ Cloud connects you to several powerful systems:
1. HPC
- JUPITER
- JUWELS
- JURECA
- JUSUF
- DEEP
For access to the HPC systems, please visit JuDoor.
2. Cloud
JSC-Cloud is accessible to everyone, providing flexible cloud resources without the need for an HPC account. For users utilizing JSC-Cloud, we offer several resource configurations, including:
- 2GB RAM, 1 CPU for 7 days
- 4GB RAM, 1 CPU for 2 days
- 8GB RAM, 2 CPUs for 10 hours
Resource Availability
These flavors may change in the future and are dependent on the available resources.
With JSC-Cloud, each user enjoys 25GB of persistent storage, allowing you to keep your important work safe and accessible. This feature is available not only in the default JupyterLab versions but also when using custom Docker images. Whether you're developing in a standard environment or tailoring your setup with your own images, you can rest assured that your data is securely stored and easily retrievable.
3. Quantum Computing
JUNIQ Cloud provides access to all the quantum computing resources available at the JUNIQ facility, including:
Quantum Processing Units (QPUs)
JUNIQ Cloud provides access to the following cutting-edge quantum computing platforms:
| QPU Platform | QPU Name (to be used in myqlm programs) | Device | Number of Qubits | Typical Use cases |
|---|---|---|---|---|
| Quantum Annealer | JUPSI* |
JUPSI Jülich Pioneer for Spin Interference • From D-Wave (Advantage2) • More than 4,400 superconducting qubits • More than 40,000 couplers • 20-way qubit connectivity (Zephyr-12 topology) |
4,400+ | Optimization problems, quantum annealing |
| Gate based - Superconducting | Jiqcer5QPU ** |
JIQCER-5 Jülich IQM Quantum Computing ExpeRience • From IQM (Spark) • 5 superconducting flux-tunable grounded transmon qubits • 4 couplers • Square lattice topology |
5 | Quantum circuit execution, gate-based algorithms |
| Analog - Neutral Atom | JadeQPU |
JADE Jülich neutral Atom DEvice • Orion system of Pasqal • More than 100 neutral rubidium atom qubits • 1D and 2D atom geometry |
100+ | Analog quantum simulation |
| Gate based - Trapped Ion | - |
JION Jülich trapped-ION quantum computer • From eleQtron • About 30 ionized ytterbium atom qubits • MAGIC (Magnetic Gradient Induced Coupling) for all-to-all qubit connectivity |
30 | Quantum circuit execution |
| Gate based - Semiconductor | - |
JAQUI-5 Jülich ArQUe semIconductor quantum computer • From ARQUE Systems • 5 electron spin qubits in silicon (Si/SiGe) • Linear Nearest Neighbor (LNN) topology via electron shuttle across a few traps |
5 | Quantum circuit execution |
| Gate based - Superconducting | QSolidQPU10 |
QSolid QPU • Developed within QSolid project • 10 qubits • Linear topology |
10 | Quantum circuit execution |
*JUPSI is currently not accessible via Qaptiva, see DWave setup for more information on its access.
Similarly the following emulators are available:
| QPU Platform | QPU Name (to be used in myqlm programs) | Max Capacity (Qubits) | Typical Use cases / Comments |
|---|---|---|---|
| Quantum simulator - Neutral Atom | PasqalQPU1** | 10 | Emulator of JADE QPU via QuTiP |
| Gate based - Trapped Ion | EmuEleqtronQPU** | 10 | Digital Twin of JION QPU by ParityQC |
| Gate based - LinAlg | LinAlg | 32 | Noiseless (perfect). Exact simulation |
| Gate based - LinAlg GPU | Linalg | 30 | Noiseless, exact, on GPU. Instantiate QPU via Linalg(use_GPU = True) |
| Gate based - MPS | MPS | Depends on circuit entanglement, >= 32 | Noiseless. Exact simulation |
| Gate based - MPO | MPO | Depends on circuit entanglement, >= 32 | Noiseless and noisy – approximate simulation |
| Gate based - QPEG | QPEG | Depends on circuit entanglement, >= 32 | Noiseless. Approximate simulation |
| Gate based - Stabilizers | Stabs | 1000+ | Noiseless, exact simulation of stabilizer circuits |
| Gate based - Feynman | Feynman | 32+ | Noiseless, exact simulation with Feynman integral path. Suited for sparse circuits only |
| Gate based - BDD | Bdd | 32+ | Noiseless, exact simulation with binary decision diagrams |
| Gate based - Noisy Deterministic | NoisyQProc | 16 | Noisy. Exact computation of density matrix |
| Gate based - Noisy Stochastic | NoisyQProc | 32 | Noisy. Sampling with Monte-Carlo technique |
| Analog - QuTip | QutipQPU | 20 | Exact analog simulation wrapped over QuTip |
| Analog - Native | AnalogQPU | 29 | Exact analog simulation (faster) |
| Analog - MPS Trajectory | MPSTraj | 80+ | Approximate analog simulation |
| Annealing - Simulated | SQAQPU | 5000 | Simulated annealing |
** These names on Qaptiva are temporary and may change in the future.
Individuals access to these resources is subjected to the project you are affiliated with. For more information on how to gain access, please refer to the access guide.
Software Stack
For quantum computing work, JUNIQ Cloud provides the PyQuantum kernel which includes:
- myQLM: Quantum Learning Machine framework used for accessing quantum resources via Qaptiva
- D-Wave Ocean: D-Wave quantum annealing tools
- Qiskit: IBM's quantum computing framework
- Other quantum libraries: Pre-installed quantum computing packages