Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/ptsolvers/adalovelace22-gpu
GPU-computing in geodynamics workshop material
https://github.com/ptsolvers/adalovelace22-gpu
geodynamics gpu julialang
Last synced: 17 days ago
JSON representation
GPU-computing in geodynamics workshop material
- Host: GitHub
- URL: https://github.com/ptsolvers/adalovelace22-gpu
- Owner: PTsolvers
- License: mit
- Created: 2022-08-25T06:39:06.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2022-09-01T14:17:58.000Z (over 2 years ago)
- Last Synced: 2023-04-01T03:37:21.098Z (almost 2 years ago)
- Topics: geodynamics, gpu, julialang
- Language: Julia
- Homepage:
- Size: 12.1 MB
- Stars: 3
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Ada Lovelace 2022 GPU
GPU-computing in geodynamics workshop material
Program: https://meetings.copernicus.org/2022AdaLovelaceWorkshop/programme/programme.html
## Automatic notebook generation
The presentation slides and the demo notebook are self-contained in a Jupyter notebook [alw22-gd-gpu.ipynb](alw22-gd-gpu.ipynb) that can be auto-generated using literate programming by deploying the [alw22-gd-gpu.jl](alw22-gd-gpu.jl) script.
To reproduce:
1. Clone this git repo
2. Open Julia and resolve/instantiate the project
```julia-repl
using Pkg
Pkg.activate(@__DIR__)
Pkg.resolve()
Pkg.instantiate()
Pkg.add(url="https://github.com/EnzymeAD/Enzyme.jl.git", rev="main")
```
_:warning: Note that the `main` branch of Enzyme.jl should be used._
3. Run the deploy script
```julia-repl
julia> using Literate
julia> include("deploy_notebooks.jl")
```
4. Then using IJulia, you can launch the notebook and get it displayed in your web browser:
```julia-repl
julia> using IJulia
julia> notebook(dir="./")
```
_To view the notebook as slide, you need to install the [RISE](https://rise.readthedocs.io/en/stable/installation.html) plugin_
## Self-contained adjoint-based GD inversion scripts
The scripts used to produce the adjoint-based gradient descent point-wise inversion for the power-law prefactor in a free-surface channel flow are accessible in the [scripts](scripts) folder:
- GPU version: [free_surface_flow_enzyme_2D_cuda.jl](scripts/free_surface_flow_enzyme_2D_cuda.jl)
- CPU version: [free_surface_flow_enzyme_2D.jl](scripts/free_surface_flow_enzyme_2D.jl)
## Resources
#### The pseudo-transient method
GMD paper: https://doi.org/10.5194/gmd-15-5757-2022
- Stokes flow examples: https://github.com/PTsolvers/PseudoTransientStokes.jl
- Diffusion solver examples: https://github.com/PTsolvers/PseudoTransientDiffusion.jl
#### Julia packages
- [ParallelStencil.jl](https://github.com/omlins/ParallelStencil.jl)
- [ImplicitGlobalGrid.jl](https://github.com/eth-cscs/ImplicitGlobalGrid.jl)
- [CUDA.jl](https://github.com/JuliaGPU/CUDA.jl)
- [Enzyme.jl](https://github.com/EnzymeAD/Enzyme.jl)
#### Courses and resources
- ETHZ course on solving PDEs with GPUs: https://pde-on-gpu.vaw.ethz.ch
- More [here](https://pde-on-gpu.vaw.ethz.ch/extras/#extra_material)
#### Misc
- Frontier GPU multi-physics solvers: https://ptsolvers.github.io/GPU4GEO/