How to manipulate multidimensional array take out each specific size and put it back?

How to crop array into specific size, in here 32x32. from left top to right? is there built-in to this ways? i read somewhere implement manual ways, make numpy slower. so i think can improve this manipulation.

import numpy

frame = numpy.full((1080, 1920), 255)
frameBlocks = numpy.array()
wBlock = 32
hBlock = 32

# Determine total block each row because sometime the latest block not exactly fit size.
wBlockLength = numpy.floor(frame.shape[1] / wBlock)
hBlockLength = numpy.floor(frame.shape[0] / hBlock)

# Crop from left to right with 32 x 32
for hBL in range(0, hBlockLength):
  hIndex = hBL*hBlock

  for wBL in range(0, wBlockLength):
    block = numpy.zeros(wBlock, hBlock)
    wIndex = wBL*wBlock

    for hB in range(0, hBlock):
      for wB in range(0, wBlock):
         block[hB][wB] = frame[hIndex][wIndex]
         wIndex += 1
    hIndex += 1

    frameBlocks.append(block)

# create index order to shuffle
blocksLength = len(frameBlocks)
index = numpy.array(range(0, blocksLength))

# Shuffle order by index
frameBlocks[index]

my other question how to turn blocks that already suffle into frame? exactly it simple just repeat that loop again. but, how about in numpy ways?