3. What kind of external libraries can I use to accelerate my code?

The Liquid Engine hinges on having multiple implementations so it can choose the fastest one. Usually, each implementation employs a different acceleration strategy or leverages different hardware devices. The Liquid Engine supports virtually any acceleration strategy you might want to employ.

In this section we list external libraries and their documentation that are commonly used in accelerating python code. Any one of this strategies or a combination thereof can be used to build a Liquid Engine compatible implementation.

• Cython - documentation here

• Numba - documentation here

• Transonic - documentation here

• Dask - documentation here

• CuPy - documentation here

• PyOpenCL - documentation here