<meta http-equiv="Content-Type" content="text/html; charset=GB18030"><div><div>Dear pwscf users</div><div>Recently I am calculating a system that needs to consider SOC, I use the Norm conservation pseudopotential with full relativistic, when I run vc-relax, I have such kind error:</div><div><br></div><div>Initial potential from superposition of free atoms</div><div>starting charge  103.98524, renormalised to  104.00000</div><div>Starting wfcs are   66 randomized atomic wfcs +   94 random wfcs</div><div>%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%</div><div>Error in routine cdiaghg (159):</div><div>eigenvectors failed to converge</div><div>%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%</div><div>stopping ...</div><div><br></div><div>But when I add:</div><div>noncolin = .true.</div><div>lspinorb = .true.</div><div>in the input file, the error will disappear and the program can run normally, however, SOC will be considered in this way in vc-relax. I notice that SOC is not supported in structural optimization. I want to know how to use the Norm conservation pseudopotential with full relativistic for structural optimization?</div><div><br></div><div>Thanks in advance,</div><div>Dongze fan</div></div><div><br></div>