Hands-On GPU Computing with Python

Interoperability between CuPy and Numba within a single Python program

In this section, we try to test the interoperability between two different modules within the same Python program, namely, CuPy and Numba. So, we import the cuda module from numba and cupy, as well:

from numba import cuda #Using Numbaimport cupy as cp #Using CuPyfrom timeit import default_timer as timerN = [e-mail] multiply(p, q):    # Thread id in a 1D block    tx = cuda.threadIdx.x    # Block id in a 1D grid    ty = cuda.blockIdx.x    # Number of threads per block    bw = cuda.blockDim.x    # Compute flattened index inside the array    index = tx + ty * bw

Like on our previous program, we compute the product based on a condition, as shown in the following code:

 if index < N: # ...