https://github.com/tortar/dynamicdiscretesamplerscomparisons
Performance Comparisons between different implementations of Dynamic Discrete Samplers
https://github.com/tortar/dynamicdiscretesamplerscomparisons
benchmark comparison dynamic-sampling julia sampling
Last synced: 8 months ago
JSON representation
Performance Comparisons between different implementations of Dynamic Discrete Samplers
- Host: GitHub
- URL: https://github.com/tortar/dynamicdiscretesamplerscomparisons
- Owner: Tortar
- License: mit
- Created: 2025-05-27T18:39:32.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-06-25T14:42:19.000Z (about 1 year ago)
- Last Synced: 2025-07-19T01:27:10.393Z (11 months ago)
- Topics: benchmark, comparison, dynamic-sampling, julia, sampling
- Language: C++
- Homepage:
- Size: 9.48 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# DynamicDiscreteSamplersComparisons
Performance and exactness comparisons between different implementations of Dynamic Discrete Samplers.
Currently, it compares four algorithms:
- EBUS (https://github.com/LilithHafner/WeightVectors.jl)
- BUS (https://github.com/CUHK-DBGroup/WSS-WIRS)
- FT (https://github.com/manpen/dynamic-weighted-index)
- DPA* (https://github.com/Daniel-Allendorf/proposal-array)
Three performance benchmarks and one exactness benchmark are performed:
- performance test on static sampling
- performance test on dynamic sampling with a fixed domain
- performance test on dynamic sampling with a variable domain
- exactness test by an exponential decay strategy
For more information, refer to the paper listed in the citation section.
## Execution
To run the benchmarks on Linux, first install the necessary softwares with
```
bash install.sh
```
and then run the benchmarks with
```
bash run.sh
```
## Results
The results are stored in csv format in the `data` folder, and as plots in the
`figures` folder:
## Citation
If you use this package in a publication, or simply want to refer to it, please cite the paper below:
```
@misc{hafner2025exactsampler,
title={An Exact and Efficient Sampler for Dynamic Discrete Distributions},
author={Lilith Orion Hafner and Adriano Meligrana},
year={2025},
eprint={2506.14062},
archivePrefix={arXiv},
primaryClass={cs.DS},
url={https://arxiv.org/abs/2506.14062},
}
```
This comparison accompanies the EBUS implementation available at https://github.com/LilithHafner/WeightVectors.jl.