Dear Lijun,
1) Depending on the models the truncation error can be very small. Specially at the edges of the system, where the Hilbert space still didn’t grow a lot.
2) The main output of the MPS codes is stored in the *out.h5 files that you can easily inspect with the ALPS Python tools. The truncation error for us is an iteration/sweep measurement that we usually don’t write in XML format.
Yes, the truncation error will normally saturate with the number of sweeps, and can be decreased by seeing more states.
Best 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 15 Apr 2016, at 08:58, Forwordom mljphy@qq.com wrote:
Dear Dolfi,
Recently, I've been using the dmrg and mps_optim two algorithms provided by ALPS 2.2.0 to calculate some properties for the same quantum systems. Then, I have compared the truncation error part of outputs from two algorithms. I have two questions about truncation error:
(1)Firstly,I've noticed that the value calculated for the truncation error comes out to be zero for some of the iterations. It seems to occur on any DMRG programs I run, even the simple ones provided in the DMRG tutorials. Is this result correct?
(2)Secondly,I can find final truncation error in *.out.xml file if using dmrg algorithm, but can not find this value when using mps_optim algorithm. How to get final truncation error for a run using mps_optim. Actually, I hope to control truncation error by adjusting the number of sweeps and the number of states kept.
Any insight and advice is much appreciated. Thanks.
Lijun Meng Xiangtan University, Hunan Pro, China