Hi Dave,<div><br></div><div>I was expecting that the code thinks of the composition is Ax before and after optimization and that's why I was surprised to see the error of incorrect composition. But perhaps it's more subtle than this.</div>
<div><br></div><div>Wei<br>
<br><br><div class="gmail_quote">On Mon, Jul 23, 2012 at 7:27 AM, David Lonie <span dir="ltr"><<a href="mailto:loniedavid@gmail.com" target="_blank">loniedavid@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="HOEnZb"><div class="h5">On Sun, Jul 22, 2012 at 5:00 PM, Wei Lai <<a href="mailto:laiwei@msu.edu">laiwei@msu.edu</a>> wrote:<br>
> I was using GULP together with XtalOpt on a compound AxByCz. I gave the<br>
> coordinates of B and C in the GULP input file and let XtalOpt to generate A<br>
> coordinates randomly. It seemed to work fine for initial structures of the<br>
> first generation. However, it gave the error of "XtalOpt::checkXtal:<br>
> Composition incorrect." at the end of the first generation and no second<br>
> generation structures could be created. I was wondering what caused this<br>
> error. Thanks.<br>
<br>
</div></div>XtalOpt expects structures to have the same composition before and<br>
after optimization, and as far as it can tell, the structure is purely<br>
Ax. Fixed atomic positions are not supported by the code, nor is<br>
supplying additional atoms directly in the input file.<br>
<br>
Hope this helps,<br>
Dave<br>
</blockquote></div><br></div>