<div>Hi,All.</div>
<div> </div>
<div>I came to a CRASH with this error information in nscf calculation :</div>
<div>"%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%<br> task # 1<br> from cgramg1_k : error # 1<br> negative or zero norm in S<br>%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% "
</div>
<div> </div>
<div>I browsed the mail list, it seems this error is related to ultrasoft pseudopotential's bad behavior, as pointing out by paolo</div>
<div>"S is the overlap matrix in the Ultrasoft pseudopotential formalism. It has to be positive definite. Likely one (or more than one) of the pseudopotentials is bad " </div>
<div>But no CRASH is occured in 'scf' or 'relax' calculation, it's too strange to belevie the problem of pseudopotential,</div>
<div>I also try to use 'cg' method for diagonalization instead of 'david ' method, and no such CRASH would be occured.</div>
<div> </div>
<div>Regards,</div>
<div>hai-ping</div>
<div> </div>
<div> </div>