mkite: distributed computing platform for high-throughput materials simulations
Project description
What is mkite?
mkite is a suite of tools for running high-throughput materials simulations in distributed computing platforms. The mkite suite decouples the production database from client workers, facilitating scaling of simulations across heterogeneous computing environments. The infrastructure enables exploration of combinatorial materials spaces using workflows, recipes, data visualizations, and more.
Some advantages of mkite:
- It stores and organizes complex materials workflows on databases. For example, mkite allows creating workflows with more than one parent input branch (say, interfaces between solids and molecules).
- The server is agnostic to the computing environments where the tasks are performed, and the clients are unaware of the production database. This facilitates distributing the tasks across heterogeneous computing systems.
- It provides textual descriptions for workflows, and enables adapting them on-the-fly. This helps as a "lab notebook" for computational materials scientists.
- It is adaptable to many software packages and inputs. The recipe system also interacts well with other libraries, such as ASE, pymatgen, cclib, and more.
Documentation
General tutorial for mkite
and its plugins are available in the main documentation.
Complete API documentation is pending.
Installation
To install mkite_db
, first install mkite_core
and mkite_engines
.
Then, install this repository with pip:
pip install mkite_core mkite_engines
pip install mkite_db
Alternatively, for a development version, clone this repo and install it in editable form:
pip install -U git+https://github.com/mkite-group/mkite_db
Contributions
Contributions to the entire mkite suite are welcomed. You can send a pull request or open an issue for this plugin or either of the packages in mkite. When doing so, please adhere to the Code of Conduct in the mkite suite.
The mkite package was created by Daniel Schwalbe-Koda dskoda@ucla.edu.
Citing mkite
If you use mkite in a publication, please cite the following paper:
@article{mkite2023,
title = {mkite: A distributed computing platform for high-throughput materials simulations},
author = {Schwalbe-Koda, Daniel},
year = {2023},
journal = {arXiv:2301.08841},
doi = {10.48550/arXiv.2301.08841},
url = {https://doi.org/10.48550/arXiv.2301.08841},
arxiv={2301.08841},
}
License
The mkite suite is distributed under the following license: Apache 2.0 WITH LLVM exception.
All new contributions must be made under this license.
SPDX: Apache-2.0, LLVM-exception
LLNL-CODE-848161
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file mkite_db-24.10.23.tar.gz
.
File metadata
- Download URL: mkite_db-24.10.23.tar.gz
- Upload date:
- Size: 100.4 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.9.20
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 7c451b09f001a3a6ca288fcbcf5c1ef5049d7bf1d98e848e4d0eb279399b440b |
|
MD5 | b45e51ff7e0aa3a4be255eeb4878dbed |
|
BLAKE2b-256 | 8f4c091725b86d9fc67b8f15183227859d61598eff7a159adc10326797da4c66 |
File details
Details for the file mkite_db-24.10.23-py3-none-any.whl
.
File metadata
- Download URL: mkite_db-24.10.23-py3-none-any.whl
- Upload date:
- Size: 86.8 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.9.20
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 7a0229878285d93dcd46b93347ccaa75c0054424e12bfdb8fc82b58f389f4aa9 |
|
MD5 | 8e155e1d797ae0c5cc01419c2a7b935b |
|
BLAKE2b-256 | d48b63383ceb8b9ed4419543fb77c00251cba0536acbce9d1e28d824b6d13fdf |