Dear Michele Dolfi,
I meet such error for two times. The system informs me that dmrg.exe has an error. The information contains "ReadDiagDataFromFile". I wonder whether it is the case: while dmrg runs, it saves some temporary datas, which may be saved to the same folder.
In theory the filenames are made unique by system calls to _mktemp().
If you really want to be sure you can set the parameter TEMP_DIRECTORY to point to different directories.
“ReadDiagDataFromFile” is again a Python function. There must be another reason why dmrg.exe gave an error.
By the way, recently I use dmrg to calculate the correlation such as b^{\dagger}_i b_j. Since there are complex parameters in the Hamiltonian, the correlation may be complex number. It seems that the present version only return the real part of the measurement. I wonder whether it is the case.
I don’t know the dmrg code so well, but indeed it seems to cast the result to real:
#530:dmrg/dmrg.h av.push_back(real(iter++->value()));
These should not be an issue with the MPS codes, which unfortunately are outdated in the last Windows binary. If you have the possibility to compile ALPS from sources (either on some Unix machine or in a virtual machine), You should definitely give it a try.
Best regards,
Michele
--
ETH Zurich
Michele Dolfi
Institute for Theoretical Physics
HIT G 32.4
Wolfgang-Pauli-Str. 27
8093 Zurich
Switzerland
+41 44 633 78 56 phone
+41 44 633 11 15 fax