Databases
StackOverflow2013
Postsdbo
POcython memoryview slower than expected
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POcython memoryview slower than expected
    primarykey
    data
    text
    <p>I've started using memoryviews in cython to access numpy arrays. One of the various advantages they have is that they are considerably faster than the old numpy buffer support: <a href="http://docs.cython.org/src/userguide/memoryviews.html#comparison-to-the-old-buffer-support" rel="nofollow">http://docs.cython.org/src/userguide/memoryviews.html#comparison-to-the-old-buffer-support</a></p> <p>However, I have an example where the old numpy buffer support is faster than memoryviews! How can this be?! I wonder if I'm using memoryviews correctly?</p> <p>This is my test:</p> <pre><code>import numpy as np cimport numpy as np cimport cython @cython.boundscheck(False) @cython.wraparound(False) cpdef np.ndarray[np.uint8_t, ndim=2] image_box1(np.ndarray[np.uint8_t, ndim=2] im, np.ndarray[np.float64_t, ndim=1] pd, int box_half_size): cdef unsigned int p0 = &lt;int&gt;(pd[0] + 0.5) cdef unsigned int p1 = &lt;int&gt;(pd[1] + 0.5) cdef unsigned int top = p1 - box_half_size cdef unsigned int left = p0 - box_half_size cdef unsigned int bottom = p1 + box_half_size cdef unsigned int right = p0 + box_half_size cdef np.ndarray[np.uint8_t, ndim=2] box = im[top:bottom, left:right] return box @cython.boundscheck(False) @cython.wraparound(False) cpdef np.uint8_t[:, ::1] image_box2(np.uint8_t[:, ::1] im, np.float64_t[:] pd, int box_half_size): cdef unsigned int p0 = &lt;int&gt;(pd[0] + 0.5) cdef unsigned int p1 = &lt;int&gt;(pd[1] + 0.5) cdef unsigned int top = p1 - box_half_size cdef unsigned int left = p0 - box_half_size cdef unsigned int bottom = p1 + box_half_size cdef unsigned int right = p0 + box_half_size cdef np.uint8_t[:, ::1] box = im[top:bottom, left:right] return box </code></pre> <p>The timing results are:</p> <p>image_box1: typed numpy: 100000 loops, best of 3: 11.2 us per loop</p> <p>image_box2: memoryview: 100000 loops, best of 3: 18.1 us per loop</p> <p>These measurements are done from IPython using %timeit image_box1(im, pd, box_half_size)</p>
    singulars
    1. PO
      singulars
      1. PTAnswer
    1. This table or related slice is empty.
    1. PTQuestion
    1. USmartinako
    1. USmartinako
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POcython memoryview slower than expected
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. POcython memoryview slower than expected
      1. This table or related slice is empty.
      1. VTUpMod
    1. COI guess you are timing these functions from python? Since then the return value is a `np.ndarray` also in the second function (I assume), which may already explain the slowdown, since making the `np.ndarray` is a bit of extra work and there is not much done here overall.
      singulars
      1. POcython memoryview slower than expected
      1. USseberg
    2. COYes, I timed these from IPython with the command: %timeit image_box1(im, pd, box_half_size) I have just edited my question to include timing from within cython. memoryviews are still slower!
      singulars
      1. POcython memoryview slower than expected
      1. USmartinako
    3. COcorrection! you are right, the delay is in the conversion from numpy array to memoryview!
      singulars
      1. POcython memoryview slower than expected
      1. USmartinako
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload