[Numpy-discussion] question about the documentation of linalg.solve

Alan G Isaac aisaac@american....
Wed Nov 19 13:27:11 CST 2008


Thanks Charles and Josh, but my question
about the documentation goal remains.

Here is how this came up.  I mentioned
to a class that I have using NumPy that
solving Ax=b with an inverse is
computationally wasteful and also has
accuracy problems, and I recommend
using `solve` instead.  So the question
arises: what algorithm does `solve` use?
I speculate Gaussian elimination, but
cannot get a quick answer from `help` nor
from the web.

In contrast, if they are using Matlab,
they look at
<URL:http://www.mathworks.com/access/helpdesk/help/techdoc/index.html?/access/helpdesk/help/techdoc/ref/arithmeticoperators.html&http://www.mathworks.com/access/helpdesk/help/techdoc/ref/inv.html>
which tell them the algorithm is Gaussian elimination
with partial pivoting and provide a link
<URL:http://www.mathworks.com/access/helpdesk/help/techdoc/index.html?/access/helpdesk/help/techdoc/ref/arithmeticoperators.html&http://www.mathworks.com/access/helpdesk/help/techdoc/ref/inv.html>
to algorithm details.

So my question is not just what is the algorithm
but also, what is the documentation goal?

Thanks,
Alan


More information about the Numpy-discussion mailing list