Dear all,
I would like to use the ALPS dmrg routines to simulate the transversal spin-1/2 XXZ Heisenberg model, i.e., with magnetic field in X-direction.
While a magnetic field in Z-direction [Hz = B*Sz(i)] poses no problems, a magnetic field in X-direction [Hx = B*(Splus(i) + Sminus(i))] destroys convergence completely.
I am aware that Hx violates Sz-conservation, but XXZ+Hz is easily solved by the dmrg package even without assuming conserved quantum numbers, so I guess conserved quantum numbers are probably not an issue for the small systems I'm considering (32 sites). Is the bad convergence in the Hx case something which is expected or do you rather suspect an error in my configuration?
Thanks a lot! Sincerely,
Thomas
Hi Thomas,
I don’t really know what happens with the `dmrg` code, I think I could reproduce the using the ALPS tutorials and adding the parameter Gamma=1. Just by looking at the local magnetization one sees a quite ugly non-symmetric curve.
I just tried to plugin the same parameters in the new MPS code `mps_optim`, and the result seems better. The parameters are attached. It is not really about understanding the problem, but more a workaround. I hope it could help you!
Regards, Michele
-- ETH Zurich Michele Dolfi Institute for Theoretical Physics HIT G 32.4 Wolfgang-Pauli-Str. 27 8093 Zurich Switzerland
dolfim@phys.ethz.ch www.itp.phys.ethz.ch
+41 44 633 78 56 phone +41 44 633 11 15 fax
On May 19, 2015, at 9:18 AM, Thomas SCHMIDT thomas.schmidt@uni.lu wrote:
Dear all,
I would like to use the ALPS dmrg routines to simulate the transversal spin-1/2 XXZ Heisenberg model, i.e., with magnetic field in X-direction.
While a magnetic field in Z-direction [Hz = B*Sz(i)] poses no problems, a magnetic field in X-direction [Hx = B*(Splus(i) + Sminus(i))] destroys convergence completely.
I am aware that Hx violates Sz-conservation, but XXZ+Hz is easily solved by the dmrg package even without assuming conserved quantum numbers, so I guess conserved quantum numbers are probably not an issue for the small systems I'm considering (32 sites). Is the bad convergence in the Hx case something which is expected or do you rather suspect an error in my configuration?
Thanks a lot! Sincerely,
Thomas
Prof. Dr. Thomas L. Schmidt Theory of Mesocopic Systems 162a, avenue de la Faïencerie Phone: (+352) 46 66 44 6158 Université du Luxembourg Email: thomas.schmidt@uni.lu L-1511 Luxembourg http://www.tlschmidt.com
Dear Michele,
thanks a lot for your help! I've tested it and 'mps_optim' seems to work very nicely for this problem.
Sincerely,
Thomas
Le 19.05.2015 11:27, Michele Dolfi a écrit :
Hi Thomas,
I don’t really know what happens with the `dmrg` code, I think I could reproduce the using the ALPS tutorials and adding the parameter Gamma=1. Just by looking at the local magnetization one sees a quite ugly non-symmetric curve.
I just tried to plugin the same parameters in the new MPS code `mps_optim`, and the result seems better. The parameters are attached. It is not really about understanding the problem, but more a workaround. I hope it could help you!
Regards, Michele
-- ETH Zurich Michele Dolfi Institute for Theoretical Physics HIT G 32.4 Wolfgang-Pauli-Str. 27 8093 Zurich Switzerland
dolfim@phys.ethz.ch mailto:dolfim@phys.ethz.ch www.itp.phys.ethz.ch http://www.itp.phys.ethz.ch
+41 44 633 78 56 phone +41 44 633 11 15 fax
On May 19, 2015, at 9:18 AM, Thomas SCHMIDT <thomas.schmidt@uni.lu mailto:thomas.schmidt@uni.lu> wrote:
Dear all,
I would like to use the ALPS dmrg routines to simulate the transversal spin-1/2 XXZ Heisenberg model, i.e., with magnetic field in X-direction.
While a magnetic field in Z-direction [Hz = B*Sz(i)] poses no problems, a magnetic field in X-direction [Hx = B*(Splus(i) + Sminus(i))] destroys convergence completely.
I am aware that Hx violates Sz-conservation, but XXZ+Hz is easily solved by the dmrg package even without assuming conserved quantum numbers, so I guess conserved quantum numbers are probably not an issue for the small systems I'm considering (32 sites). Is the bad convergence in the Hx case something which is expected or do you rather suspect an error in my configuration?
Thanks a lot! Sincerely,
Thomas
Prof. Dr. Thomas L. Schmidt Theory of Mesocopic Systems 162a, avenue de la Faïencerie Phone: (+352) 46 66 44 6158 Université du Luxembourg Email: thomas.schmidt@uni.lu mailto:thomas.schmidt@uni.lu L-1511 Luxembourg http://www.tlschmidt.com
comp-phys-alps-users@lists.phys.ethz.ch