Modern instructions and GPU support

NumPower was designed from the ground up to utilize AVX2, BLAS and the GPU to further improve performance.

N-dimensional arrays

Inspired by NumPy, NumPower arrays are N-dimensional and have efficient methods for slicing and indexing N-dimensional arrays.

Mathematical Tools

Count on dozens of high-performance mathematical operations, including: trigonometry, statistics, linear algebra and more.

CUDA Support

NumPower was developed from scratch to operate with GPUs that support CUDA. You can easily store and perform operations using your GPU's memory and processor.

Image processing

You can use GD images as input to perform various image processing. You can use slicing and convolution methods for example to crop or apply custom filters.

Efficient memory usage

Initially implemented with single precision floats, NDArray allows you to work with more data with less memory. We plan to add more types in the near future.