Why does a Numpy with Intel MKL have the same performance as a normal one?

[AlekseyPython]

Intel i5-4690, Kubuntu 20.10

I installed the Numpy in the usual way:

pip3 install numpy

, and in Eclipse IDE ran the matrix multiplication test:

import numpy as np
from time import time

data1 = np.random.rand(24000, 4000)
data2 = np.random.rand(4000, 16000)
t_start = time()
data3 = np.matmul(data1, data2)
t_end = time()
print('Time elapsed: ' + str(t_end-t_start))

Then I ran the same test in the Spyder IDE, that was installed with Anaconda. As you know, packets collected in Anaconda automatic use Intel MKL.

The speed of matrix multiplication is the same! Why?