https://github.com/santisoler/simpeg-benchmarks-gravity
https://github.com/santisoler/simpeg-benchmarks-gravity
Last synced: 5 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/santisoler/simpeg-benchmarks-gravity
- Owner: santisoler
- License: mit
- Created: 2023-10-25T00:11:52.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2024-08-21T18:52:56.000Z (over 1 year ago)
- Last Synced: 2024-10-16T04:21:10.565Z (about 1 year ago)
- Language: Jupyter Notebook
- Size: 5.35 MB
- Stars: 5
- Watchers: 1
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Benchmarks for comparing gravity simulations in SimPEG
## About
This repository was created to run benchmarks between the gravity forward model
available in [SimPEG][simpeg] [v0.20][simpeg-v0.20] and the new implementation
of it that was merged in
[simpeg@eb07ef133a6c5e1714ff1af181147cb172eba97c](https://github.com/simpeg/simpeg/commit/eb07ef133a6c5e1714ff1af181147cb172eba97c).
This new implementation
makes use of [Numba][numba] and the gravity kernels implemented in
[Choclo][choclo] to achieve faster run times and handle memory in a more
efficient way.
The benchmarks were set up in the Python files located in the `notebooks`
folder. They compare the two implementations in different scenarios, varying
the number of cells in the mesh, the number of receivers, enabling and
disabling parallelization (using both `mutliprocessing` and `dask` for the
SimPEG v0.20 implementation), and varying the number of threads assigned to
the simulation.
The results were saved as netCDF files (that can be opened with
[Xarray][xarray]) in the `raw` folder. Figures of these results can be found in
the `figs` folder.
## Get started
In order to run these benchmarks, you need to have a Python distribution like
[Miniforge][miniforge] installed.
Then, clone this repository:
```bash
git clone https://github.com/santisoler/simpeg-benchmarks-gravity
cd simpeg-benchmarks-gravity
```
And create a `conda` environment with all the required dependencies for running
these benchmarks:
```bash
mamba env create -f environment.yml
```
> **Important**
> I recommend using `mamba` over `conda` due to its faster speeds for resolving
> dependencies and installing packages. In case you don't have `mamba` installed,
> feel free to replace `mamba` with `conda`.
## Run the benchmarks
All benchmarks can be run by executing the Python scripts in `notebooks`
folder, and through the `benchmark-memory.sh` script.
Alternatively, we can run all benchmarks by executing the `run.sh` shell
script:
```bash
bash run.sh
```
> **Important**
> Most of the benchmarks were designed to be run on a machine with 125 GB of
> ram and a minimum of 30 threads. If your system doesn't meet these specs, you
> can modify the scripts to adjust them to your needs.
>
> The benchmarks for the "large problem" require more memory: up to ~800 GB.
[miniforge]: https://github.com/conda-forge/miniforge
[simpeg]: https://simpeg.xyz
[simpeg-v0.20]: https://github.com/simpeg/simpeg/releases/tag/v0.20.0
[numba]: https://numba.pydata.org
[choclo]: https://www.fatiando.org/choclo
[xarray]: https://docs.xarray.dev