qsimulate-open / bagel

Brilliantly Advanced General Electronic-structure Library
GNU General Public License v3.0
96 stars 43 forks source link

Problem with "algorithm" : "noopt" option in CASSCF #209

Closed tpitesa closed 2 years ago

tpitesa commented 4 years ago

Dear all,

I tried to perform the CASCI calculation by calling CASSCF with no orbital optimization (I cannot call RASCI or FCI because I want to proceed with such obtained reference states to the XMS-CASPT2 calculation and SMITH allows only HF or CASSCF reference) and the program stops immediately after printing CI vectors (seg. fault) without any error message in output file. It happens regardless the number of states, active space, molecule... Any ideas why this problem occurs? It seems like a bug in "algorithm" : "noopt" option to me or maybe I did something wrong in installation? Thanks a lot!

Best, Tomislav

shiozaki commented 4 years ago

Please post a self-contained input with a small molecule here so I can reproduce.

On Tue, Jul 28, 2020 at 8:43 AM tpitesa notifications@github.com wrote:

Dear all,

I tried to perform the CASCI calculation by calling CASSCF with no orbital optimization (I cannot call RASCI or FCI because I want to proceed with such obtained reference states to the XMS-CASPT2 calculation and SMITH allows only HF or CASSCF reference) and the program stops immediately after printing CI vectors (seg. fault) without any error message in output file. It happens regardless the number of states, active space, molecule... Any ideas why this problem occurs? It seems like a bug in "algorithm" : "noopt" option to me or maybe I did something wrong in installation? Thanks a lot!

Best, Tomislav

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nubakery/bagel/issues/209, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAKDMIVAYCVTSF7GK3NNOYTR53BYFANCNFSM4PKPKW3A .

tpitesa commented 4 years ago

Here you are:

{ "bagel" : [

{ "title" : "molecule", "basis" : "aug-cc-pVDZ", "df_basis" : "cc-pVDZ-jkfit", "angstrom" : true, "geometry" : [ { "atom" : "C", "xyz" : [ -0.7034280, -1.1405244, 0.0000000 ] }, { "atom" : "N", "xyz" : [ -1.4280570, 0.0000000, 0.0000000 ] }, { "atom" : "C", "xyz" : [ -0.7034280, 1.1405244, 0.0000000 ] }, { "atom" : "C", "xyz" : [ 0.7034280, 1.1405244, -0.0000000 ] }, { "atom" : "C", "xyz" : [ 0.7034280, -1.1405244, -0.0000000 ] }, { "atom" : "H", "xyz" : [ -1.2628864, 2.0811297, -0.0000000 ] }, { "atom" : "H", "xyz" : [ 1.2628864, 2.0811297, 0.0000000 ] }, { "atom" : "H", "xyz" : [ 1.2628864, -2.0811297, 0.0000000 ] }, { "atom" : "H", "xyz" : [ -1.2628864, -2.0811297, -0.0000000 ] }, { "atom" : "N", "xyz" : [ 1.4280570, -0.0000000, 0.0000000 ] } ] },

{ "title" : "hf" },

{ "title" : "casscf", "algorithm" : "noopt", "maxiter" :1500, "nact" : 8, "nclosed" : 16, "nstate" : 2 }

]}

shiozaki commented 4 years ago

I will fix this when I have time, but in the meantime, add an option "canonical" : true in the casscf/noopt block and the calculation will go through.

tpitesa commented 4 years ago

Thanks, this works