https://github.com/pennylaneai/lightning-on-hpc
"Hybrid quantum programming with PennyLane Lightning on HPC platforms" accompanying data and workloads
https://github.com/pennylaneai/lightning-on-hpc
cpp20 cuda gpu hpc mpi openmp python quantum quantum-computing rocm supercomputing
Last synced: 4 months ago
JSON representation
"Hybrid quantum programming with PennyLane Lightning on HPC platforms" accompanying data and workloads
- Host: GitHub
- URL: https://github.com/pennylaneai/lightning-on-hpc
- Owner: PennyLaneAI
- License: apache-2.0
- Created: 2024-03-01T18:59:24.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-07-09T19:05:28.000Z (about 1 year ago)
- Last Synced: 2024-12-23T00:05:20.177Z (10 months ago)
- Topics: cpp20, cuda, gpu, hpc, mpi, openmp, python, quantum, quantum-computing, rocm, supercomputing
- Language: Jupyter Notebook
- Homepage:
- Size: 573 KB
- Stars: 9
- Watchers: 7
- Forks: 0
- Open Issues: 8
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Hybrid quantum programming with PennyLane Lightning on HPC platforms
This repository serves accompanying data and documentation for the manuscript ["Hybrid quantum programming with PennyLane Lightning on HPC platforms"](https://arxiv.org/abs/2403.02512)
The structures of this repository is as follows:
```
DataCollection/
├─ distributed/
| ├─ LUMI_LKOKKOS_VQE
| | # Data, notebooks and scripts used for the distributed VQE workload
| | # running on 8 - 256 AMD MI250X GPUs on LUMI, using Lightning-Kokkos,
| | # OpenMPI+UCX and mpi4py.
| |
| ├─ Perlmutter_LGPU_MPI/
| | | # All materials related to fully-distributed Lightning-GPU workloads.
| | ├─ sampling_results/
| | | # Data and scripts for the 33-41 qubit sampling workload running on
| | | # 4 - 512 NVIDIA A100 (80GB) GPUs on Perlmutter, using Lightning-GPU
| | | # Cray-MPICH 8.1.25, mpi4py and CUDA 11.7.
| | |
| | ├─ sel_gradient_results/
| | | # Data and scripts for the 33-37 qubit distributed quantum gradient workload
| | | # running on 16 - 512 NVIDIA A100 (80GB) GPUs on Perlmutter, using
| | | # Lightning-GPU, Cray-MPICH 8.1.25, mpi4py and CUDA 11.7.
| | |
| ├─ Perlmutter_LGPU_QAOA/
| | | # All materials related to task-based circuit cutting Lightning-GPU workloads.
| | ├─ QAOA_QCUT_BATCHED_CUDA12/
| | | # Data and scripts for the 79-qubit circuit cutting workload, with 40k+
| | | # intermediate circuits, using batched execution, and running on between
| | | # 4 - 256 NVIDIA A100 (80GB) GPUs on Perlmutter, using Lightning-GPU
| | | # CUDA 12.2, and Ray.
| | |
| | ├─ QAOA_QCUT_ONDEMAND_CUDA12/
| | | # Data and scripts for the 79-qubit circuit cutting workload, with 40k+
| | | # intermediate circuits, using on-demand execution, and running on between
| | | # 4 - 256 NVIDIA A100 (80GB) GPUs on Perlmutter, using Lightning-GPU
| | | # CUDA 12.2, and Ray.
| | |
├─ single_node/
| ├─ GateKernels_LQUBIT/
| | | # Targeted kernel evaluation for LM, AVX2 and AVX512 implementations, with
| | | # comparisons against other HPC-focused simulators.
| ├─ VQE/
| | | # Comparison of Lightning suite across CPU (Lightning-Qubit, Lightning-Kokkos+OpenMP on AMD 3rd gen. Epyc)
| | | # and GPU (Lightning-GPU and Lightning-Kokkos+CUDA on NVIDIA A100) backends, using
| | | # VQE and a range of molecules from PennyLane's datasets.
```
The included `requirements.txt` contains all used dependencies for running the attached notebooks and plotting scripts.