<br><br><div class="gmail_quote">On Fri, May 20, 2011 at 6:33 PM, Pauli Virtanen <span dir="ltr"><<a href="mailto:pav@iki.fi">pav@iki.fi</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
On Fri, 20 May 2011 17:43:54 -0600, Charles R Harris wrote:<br>
[clip: expi]<br>
<div class="im">> Curious. I don't see this on ubuntu, it looks like a factor of 2*pi*1j<br>
> is getting added on. Looks like a corner case involving complex log or<br>
> some such.<br>
<br>
</div>The implementation is based on calculating E1(-z), and cancelling out<br>
the i*pis with a logarithm. It's a bit of a bad form on the real axis,<br>
and apparently the branch cuts are not consistently chosen with MKL.<br>
I'd guess the values will be OK away from the real axis, though,<br>
but some rewriting probably needs to be done.<br>
<font color="#888888"><br></font></blockquote><div><br>IIRC, the default flags for compiling with MKL on Windows changed between 1.5 and 1.6. I wonder if that has anything to do with this?<br><br>Chuck <br></div><br></div>