If you specify a bond type as in <BONDTERM type="1" ...> then you cannot use the # symbol in the coupling t# but need to write explicitly, e.g. t1 Matthias On Feb 20, 2013, at 3:50 PM, Alex Petrescu <tpetresc@gmail.com> wrote:

Hi Matthias,

Thanks for your reply. Below is the simplest implementation that I could cook up with free spinless fermions on two sites with *directed* bonds. The output of sparsediag, including the error, is -------------------------------------------------------------------------- parsing task files ... Starting task 1. Quantumnumber N going from 0 to 2 with increment 1 Cannot evaluate expression t# * 1 * 1 --------------------------------------------------------------------------

My short source is attached below. I hope that it's a quick task to figure out what is wrong with the implementation.

Thanks in advance, Alex

Alexandru Petrescu / Physics Department, Yale University

Code attachment

In <alps-dir>/lib/xml/lattices.xml I added ------------------------------------------------------------------------ <GRAPH name="my 2-site" vertices="2"> <EDGE source="1" target="2" type="1"/> <EDGE source="2" target="1" type="2"/> </GRAPH> -------------------------------------------------------------------------

In <alps-dir>/lib/xml/models.xml, ------------------------------------------------------------------------ <HAMILTONIAN name="free spinless fermions with bondterm type"> <PARAMETER name="mu" default="0"/> <PARAMETER name="t" default="1"/> <BASIS ref="spinless fermion"/> <SITETERM site="i"> <PARAMETER name="mu#" default="mu"/> -mu#*n(i) </SITETERM> <BONDTERM type="1" source="i" target="j"> <PARAMETER name="t#" default="t"/> -t#*(cdag(i)*c(j)+cdag(j)*c(i)) </BONDTERM> <BONDTERM type="2" source="i" target="j"> <PARAMETER name="t#" default="t"/> -t#*(cdag(i)*c(j)+cdag(j)*c(i)) </BONDTERM> </HAMILTONIAN> ------------------------------------------------------------------------- The parameter file is

LATTICE="my 2-site"; ------------------------------------------------------------------------- MODEL="free spinless fermions with bondterm type" NUMBER_EIGENVALUES = 1 CONSERVED_QUANTUMNUMBERS="N" MEASURE_ENERGY = true

t = 3.0; V = 0.0;

{mu=0.00} --------------------------------------------------------------------------

If instead "spinless fermions" with V=0 are used in the parameter file, sparsediag works, but all energies come out to be 0.