A collection of diagnostic and testing scripts for validating computational environments used in quantitative economics research and teaching.
This repository contains testing utilities to verify that computational environments are properly configured for running QuantEcon materials, including GPU-accelerated computations, neural network training, and other computational economics tasks.
Script: test_jax_gpu.py
Comprehensive diagnostic tool for NVIDIA GPU environments running JAX programs. Validates drivers, CUDA, CUDNN, and JAX GPU functionality with actual computations.
Quick usage:
cd gpu-testing
python3 test_jax_gpu.pyTests:
- NVIDIA driver installation and version
- CUDA toolkit and CUDNN library versions
- GPU hardware detection across frameworks
- JAX GPU backend functionality with performance benchmarks
Contributions are welcome! If you have additional test scripts or improvements to existing ones, please submit a pull request.
When adding new tools:
- Create a descriptive folder for the tool
- Include a comprehensive README in the folder
- Update this main README with a brief description and link
For issues related to:
- Test scripts: Open an issue in this repository
- QuantEcon materials: See QuantEcon Discourse
QuantEcon Team
MIT License
Note: This repository supports computational economics research and education using modern numerical methods and hardware acceleration.