Parallelized Quantum Programs
In distributed quantum computing, parallelism is essential for scalability and performance. Divi is designed to automatically parallelize hybrid quantum-classical programs by analyzing the quantum algorithms, partitioning the problem, and generating batch jobs.
This page provides a technical overview of how Divi handles parallelization.
Why Parallelization Matters
Quantum algorithms often require many repeated circuit executions—either for sampling, parameter optimization, or distributed subproblem evaluation. However, the number of available qubits on near-term quantum hardware is limited. Divi addresses this constraint by decomposing large optimization problems into smaller subproblems, which can be processed in parallel.
Quantum tasks are often constrained by execution latency and queue times—particularly when running on cloud-accessed quantum hardware. By parallelizing both quantum and classical portions of the workflow, Divi:
- Reduces end-to-end runtime by saturating quantum service pipelines.
- Overlaps classical computation (e.g. parameter evaluation, gradient computation) with quantum execution.
- Enables batch-style processing of parameter sets, improving throughput.
How Divi Parallelizes Quantum Workloads
Divi automates the entire workflow of splitting a principal problem into independently executable components. Here’s how the process works:
- We provide a collection of applications that have a structure that allows them to be parallelized. A user selects their problem and their parameters.
- Divi uses state-of-the-art parallelization for that problem to generate collections of circuits to execute in batch. This differs problem to problem, but generally, this introduces many circuits to run.
- Divi generates batches of circuits to run. The execution of the batch circuits is performed using distributed systems through Qoro’s cloud infrastructure through an API.
- Part of Divi’s functionality is to track the execution of the jobs and retrieving the results when complete. When the results arrive, Divi handles the aggregation step and performs optimization steps automatically.
- Once the program is complete, Divi includes visualization tools to analyze the results.
Parallel Optimization
Divi is built for performance at scale. To accelerate optimization of quantum-classical hybrid workloads, Divi supports parallel execution of subproblems.
-
Parallel Monte Carlo sampling: Randomized parameter samples are used in bulk, allowing rapid exploration of the search space. This is ideal for finding promising regions in complex or noisy cost landscapes
-
Local Optimizers with Update Tracking: Methods like Nelder-Mead or L-BFGS-B can be applied in a serial manner, with Divi managing iterative update steps and intermediate results. This allows hybrid strategies where a broad initial search is followed by local refinement.
Quantum tasks are often bottlenecked by access latency. Parallelization helps saturate quantum service pipelines while overlapping classical computation.
Divi’s Built-in Programs
We are continuously expanding our portfolio of programs. These are the ones currently implemented.
| Use Case | Main Parallelization | Main use |
|---|---|---|
| Variational Quantum Eigensolver (VQE) | Hamiltonian splitting | Quantum chemistry |
| Quantum Approximate Optimization Algorithm (QAOA) | Contraint Graph partitioning | Combinatorial optimization |
| Maximum-Likelihood Amplitude Estimation (MLAE) | Number of oracle evaluations | Monte-Carlo Integration |