Disable Cache

Support
1

I tried to disable the cache but it didn’t work, why is that?
The code is as follows:

from pynq import allocate
import numpy as np
Convblock_w_1 = allocate(shape=(16, 4, 3, 3), dtype=np.int16, cacheable=False)
Convblock_out_1 = allocate(shape=(16, 432, 432), dtype=np.int16, cacheable=False)
print(“Convblock_w_1 cacheable:”, Convblock_w_1.cacheable)
print(“Convblock_out_1 cacheable:”, Convblock_out_1.cacheable)

The printed results are as follows:

Convblock_w_1 cacheable: True
Convblock_out_1 cacheable: True

2

My pynq version is the latest 3.0.1.

I’d be very grateful for any help.

3

Hi,
I do not know why your command seems not to work. However, your command is not necessary as by default, pynq.allocate creates non-cacheable buffers. It is explained somewhere here:

FPGA Data Movement Using NumPy