HybridCloudSim – Hybrid Quantum–HPC Cloud Simulation Framework

QCloudSim is a Python-based simulation framework for modeling, executing, and analyzing hybrid quantum–HPC workloads.
It enables researchers to evaluate scheduling strategies, resource allocation policies, and device utilization trends in environments that combine noisy Quantum Processing Units (QPUs) with classical High-Performance Computing (HPC) resources such as CPUs and GPUs.

The framework supports:

• Heterogeneous resources: QPUs, CPUs, and memory bandwidth modeling.
• Hybrid workflows: Alternating quantum and classical computation stages.
• Custom scheduling: Parallel, sequential, or user-defined scheduling algorithms.
• Noise-aware modeling: Fidelity and noise considerations for realistic QPU behavior.
• Detailed logging & visualization: Gantt charts, utilization time series, and average resource usage.

---

📦 Features

• Configurable Devices

  • Built-in presets for quantum devices (e.g., IBM_Kawasaki, IBM_Kyiv).
  • Built-in presets for classical CPUs (e.g., AMD EPYC 9654, NVIDIA Grace Hopper).
  • Adjustable qubit counts, CPU cores, and memory bandwidth capacities.

• Workload Management

  • Supports job dispatch from CSV files or generated workloads.
  • Tracks arrival, start, finish times for each device stage.
  • Measures per-resource utilization over time.

• Visualization Tools

  • Gantt Charts – visualize execution phases across devices.
  • Utilization Time Series – track CPU, QPU, and memory bandwidth usage.
  • Average Utilization Bar Charts – compare overall resource demands.

---

🛠 Installation

Clone the repository:

git clone https://github.com/yourusername/QCloudSim.git
cd QCloudSim

Install dependencies

pip install -r requirements.txt

Run Use‑Case scripts

python3 main.py

or use jupyter notebook main.ipynb

Below is a minimal example of running a hybrid simulation with two QPUs and two CPUs.

from QCloud import *

PRINTLOG = False

# Devices
ibm_kawasaki = IBM_Kawasaki(env=None, name="QPU-1", printlog=PRINTLOG)
ibm_kyiv     = IBM_Kyiv(env=None, name="QPU-2", printlog=PRINTLOG)
cpu1         = CPU("CPU-1", env=None)
cpu2         = CPU("CPU-2", env=None)

# Hybrid environment
sim_env = HybridCloudSimEnv(
    qpu_devices=[ibm_kawasaki, ibm_kyiv],
    cpu_devices=[cpu1, cpu2],
    broker_class=ParallelBroker,
    job_feed_method='dispatcher',
    file_path='synth_job_batches/10-job.csv', 
    job_generation_model=None, 
    printlog=PRINTLOG
)

# Run the simulation
sim_env.run()

Repository Structure

QCloudSim/
│
├── QCloud/                 # Core framework package
│   ├── devices.py          # Device definitions (QPU, CPU, etc.)
│   ├── dependencies.py     # Presets and constants
│   ├── __init__.py         # Unified imports
│   ├── hybridcloudsimenv.py# Hybrid simulation environment
│   ├── brokers.py          # Scheduling/broker logic
│   └── utils.py            # Utility functions
│
├── synth_job_batches/      # Example job CSV files
├── Images/                 # Example plots
├── requirements.txt        # Python dependencies
└── README.md               # This file