Installation¶
This guide covers installation of Sparse Kappa and its dependencies.
Prerequisites¶
Required¶
Python 3.8 or later
NVIDIA GPU with compute capability 6.0 or higher
CUDA Toolkit 11.0 or later (11.x or 12.x)
Recommended¶
conda or virtualenv for environment management
8GB+ GPU memory for large matrices
Check Your System¶
CUDA Version¶
Check your CUDA version:
nvcc --version
If you don’t have CUDA installed, download it from NVIDIA’s website.
GPU Information¶
Check your GPU:
nvidia-smi
Installing PyTorch¶
PyTorch is the main dependency. Install the version matching your CUDA:
For CUDA 11.x¶
pip install torch
For CUDA 12.x¶
pip install torch
Verify Installation¶
from sparse_kappa.backend import torch_api as cp
print(f"PyTorch version: {cp.__version__}")
print(f"CUDA version: {cp.cuda.runtime.runtimeGetVersion()}")
# Test GPU
x = cp.array([1, 2, 3])
print(f"GPU test: {cp.sum(x)}")
Installing Sparse Kappa¶
From Source (Recommended)¶
git clone https://github.com/inEXASCALE/sparse-kappa.git
cd sparse-kappa
pip install -e .
This installs in editable mode for development.
From PyPI (Coming Soon)¶
pip install sparse-kappa
Development Installation¶
For development with testing tools:
git clone https://github.com/inEXASCALE/sparse-kappa.git
cd sparse-kappa
pip install -e ".[dev]"
This installs additional packages:
pytest - for running tests
pytest-cov - for coverage reports
black - for code formatting
flake8 - for linting
Verify Installation¶
from sparse_kappa import cond_estimate
from sparse_kappa.backend import sparse as sp
# Create test matrix
A = sp.random(100, 100, density=0.1, format='csr')
# Estimate condition number
cond = cond_estimate(A)
print(f"Condition number: {cond:.2e}")
# Check version
import sparse_kappa
print(f"Sparse Kappa version: {sparse_kappa.__version__}")
Common Issues¶
Issue: “No module named ‘PyTorch’”¶
Solution: Install PyTorch matching your CUDA version
pip install torch
Issue: “CUDA driver version is insufficient”¶
Solution: Update your NVIDIA driver
Check required driver version for your CUDA
Download from NVIDIA Driver Downloads
Install and reboot
Issue: ImportError with cuSOLVER¶
Solution: This is expected - the library handles this gracefully and falls back to alternative methods.
Issue: Out of memory¶
Solution:
Use iterative solvers instead of LU
Reduce matrix size for testing
Use methods with lower memory footprint
# Instead of LU
cond_estimate(A, norm=1, method='hager-higham', solver='lsmr')
Next Steps¶
Quick Start Guide - Learn basic usage
User Guide - Comprehensive guide
Examples - Code examples