<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title></title>
</head>
<body>
<div name="messageBodySection" style="font-size: 14px; font-family: -apple-system, BlinkMacSystemFont, sans-serif;">Hi Warren,
<div><br /></div>
<div>ogrid doesn’t solve my problem. Note that my code returns arrays that would evaluate as equal to the mgrid output. It’s just that they are copied in mgrid into a giant array, instead of broadcast:</div>
<div><br /></div>
<div><br /></div>
<div>
<div><i>In [176]: a0, b0 = np.mgrid[:5, :5]</i></div>
<div><i><br /></i></div>
<div><i>In [177]: a1, b1 = th.broadcast_mgrid((np.arange(5), np.arange(5)))</i></div>
<div><i><br /></i></div>
<div><i>In [178]: a0</i></div>
<div><i>Out[178]: </i></div>
<div><i>array([[0, 0, 0, 0, 0],</i></div>
<div><i> [1, 1, 1, 1, 1],</i></div>
<div><i> [2, 2, 2, 2, 2],</i></div>
<div><i> [3, 3, 3, 3, 3],</i></div>
<div><i> [4, 4, 4, 4, 4]])</i></div>
<div><i><br /></i></div>
<div><i>In [179]: a1</i></div>
<div><i>Out[179]: </i></div>
<div><i>array([[0, 0, 0, 0, 0],</i></div>
<div><i> [1, 1, 1, 1, 1],</i></div>
<div><i> [2, 2, 2, 2, 2],</i></div>
<div><i> [3, 3, 3, 3, 3],</i></div>
<div><i> [4, 4, 4, 4, 4]])</i></div>
<div><i><br /></i></div>
<div><i>In [180]: a0.strides</i></div>
<div><i>Out[180]: (40, 8)</i></div>
<div><i><br /></i></div>
<div><i>In [181]: a1.strides</i></div>
<div><i>Out[181]: (8, 0)</i></div>
</div>
<div><br /></div>
<div><br /></div>
</div>
<div name="messageReplySection" style="font-size: 14px; font-family: -apple-system, BlinkMacSystemFont, sans-serif;"><br />
On 9 Mar 2017, 2:05 PM +1100, Warren Weckesser <warren.weckesser@gmail.com>, wrote:<br />
<blockquote type="cite" style="margin: 5px 5px; padding-left: 10px; border-left: thin solid #1abc9c;">
<div dir="ltr"><br />
<div class="gmail_extra"><br />
<div class="gmail_quote">On Wed, Mar 8, 2017 at 9:48 PM, Juan Nunez-Iglesias <span dir="ltr"><<a href="mailto:jni.soma@gmail.com" target="_blank">jni.soma@gmail.com</a>></span> wrote:<br />
<blockquote class="gmail_quote" style="margin: 5px 5px; padding-left: 10px; border-left: thin solid #e67e22;">
<div>
<div name="messageBodySection" style="font-size:14px;font-family:-apple-system,blinkmacsystemfont,sans-serif">I was a bit surprised to discover that both meshgrid nor mgrid return fully instantiated arrays, when simple broadcasting (ie with stride=0 for other axes) is functionally identical and happens much, much faster.
<div><br /></div>
</div>
</div>
</blockquote>
<div><br />
<br /></div>
<div>Take a look at ogrid: <a href="https://docs.scipy.org/doc/numpy/reference/generated/numpy.ogrid.html">https://docs.scipy.org/doc/numpy/reference/generated/numpy.ogrid.html</a><br />
<br /></div>
<div>Warren<br />
<br />
<br /></div>
<blockquote class="gmail_quote" style="margin: 5px 5px; padding-left: 10px; border-left: thin solid #e67e22;">
<div>
<div name="messageBodySection" style="font-size:14px;font-family:-apple-system,blinkmacsystemfont,sans-serif">
<div>I wrote my own function to do this:</div>
<div><br /></div>
<div><br /></div>
<div>
<div><i>def broadcast_mgrid(arrays):</i></div>
<div><i> shape = tuple(map(len, arrays))</i></div>
<div><i> ndim = len(shape)</i></div>
<div><i> result = []</i></div>
<div><i> for i, arr in enumerate(arrays, start=1):</i></div>
<div><i> reshaped = np.broadcast_to(arr[[...] + [np.newaxis] * (ndim - i)],</i></div>
<div><i> shape)</i></div>
<div><i> result.append(reshaped)</i></div>
<div><i> return result</i></div>
</div>
<div><br /></div>
<div><br /></div>
<div>For even a modest-sized 512 x 512 grid, this version is close to 100x faster:</div>
<div><br /></div>
<div><br /></div>
<div>
<div><i>In [154]: %timeit th.broadcast_mgrid((np.arange(<wbr />512), np.arange(512)))</i></div>
<div><i>10000 loops, best of 3: 25.9 µs per loop</i></div>
</div>
<div><i><br /></i></div>
<div>
<div><i>In [156]: %timeit np.meshgrid(np.arange(512), np.arange(512))</i></div>
<div><i>100 loops, best of 3: 2.02 ms per loop</i></div>
<div><i><br /></i></div>
<div><i>In [157]: %timeit np.mgrid[:512, :512]</i></div>
<div><i>100 loops, best of 3: 4.84 ms per loop</i></div>
</div>
<div><br /></div>
<div><br /></div>
<div>Is there a conscious design decision as to why this isn’t what meshgrid/mgrid do already? Or would a PR be welcome to do this?</div>
<div><br /></div>
<div>Thanks,</div>
<div><br /></div>
<div>Juan.</div>
</div>
</div>
<br />
______________________________<wbr />_________________<br />
NumPy-Discussion mailing list<br />
<a href="mailto:NumPy-Discussion@scipy.org">NumPy-Discussion@scipy.org</a><br />
<a href="https://mail.scipy.org/mailman/listinfo/numpy-discussion" rel="noreferrer" target="_blank">https://mail.scipy.org/<wbr />mailman/listinfo/numpy-<wbr />discussion</a><br />
<br /></blockquote>
</div>
<br /></div>
</div>
_______________________________________________<br />
NumPy-Discussion mailing list<br />
NumPy-Discussion@scipy.org<br />
https://mail.scipy.org/mailman/listinfo/numpy-discussion<br /></blockquote>
</div>
</body>
</html>