Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/nixigaj/nbody-proj
University assignment | Programatically simulating bodies interaction in space
https://github.com/nixigaj/nbody-proj
Last synced: 1 day ago
JSON representation
University assignment | Programatically simulating bodies interaction in space
- Host: GitHub
- URL: https://github.com/nixigaj/nbody-proj
- Owner: nixigaj
- Created: 2024-09-05T12:17:52.000Z (4 months ago)
- Default Branch: master
- Last Pushed: 2024-09-17T16:13:57.000Z (4 months ago)
- Last Synced: 2024-09-18T02:00:55.290Z (4 months ago)
- Language: Python
- Homepage: https://uppsala.instructure.com/courses/97493/assignments/275664
- Size: 1.08 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# nbody-proj
Erik Junsved and Felix Ljungkvist
Programatically simulating bodies interaction in space.
## Usage
This program is only tested on Glibc-based Linux distributions.
### Dependencies
- Python
- Pip
- Make
- GCC (optional, for helper tools)
- Go (optional, for helper tools)
### Running simulation
Install Python dependencies:
```sh
pip install -r requirements.txt
```
Execute the program with `./src/main.py`
```
Usage: ./src/main.py calc [no_iterations] [input_file] [output_file] [full_sim_output_file (optional)]
./src/main.py show [no_iterations] [full_sim_input_file]
```
### Building helper tools
```sh
make
```
The executables for the helper tools are now available as `./cmp_gal` for comparing GAL files
and `./print` for printing the contents of GAL files in a human readable fashion.
## Results
### Machine specifications
- **OS:** Red Hat Enterprise Linux 9.3
- **Kernel version:** 5.14.0
- **Python/Pip version:** 3.12.1
- **CPU:** AMD Ryzen 5 5600 at 3.5 GHz (6 cores, 12 threads, AVX2 support)
- **RAM:** 64GB Dual-channel ECC DDR4 at 3000 MT/s
### Data table for 3000 bodies and 100 steps
| Test | Real (s) | User (s) | Kernel (s) |
|---------------------------|----------|----------|------------|
| Original | 3099.676 | 3091.845 | 1.010 |
| Key-value array access | 7046.221 | 7027.264 | 2.022 |
| Traditional optimizations | 4018.918 | 4008.826 | 0.918 |
| Numba JIT for force | 3.181 | 4.556 | 0.136 |
Note that the original force function could not be JIT compiled by Numba.
### Data table for 10000 bodies and 1000 steps
| Test | Real (s) | User (s) | Kernel (s) |
|---------------------|----------|----------|------------|
| Numba JIT for force | 220.46 | 221.173 | 0.300 |