For t=0 this is just a classical Ising model and you should thus use a classical MC code, for example the ALPS spinmc code.

Matthias Troyer

On 2 May 2012, at 08:20, Jianming Cai wrote:

Dear Prof. Troyer,

I use dir_loopsee to simulate the hardcore boson model on a triangular lattice as follows:


<HAMILTONIAN name="hardcore boson">
  <PARAMETER name="V0" default="1"/>
  <PARAMETER name="V1" default="1"/>
  <PARAMETER name="V2" default="1"/>
  <PARAMETER name="t" default="0"/>
  <PARAMETER name="t0" default="t"/>
  <PARAMETER name="t1" default="t"/>
  <PARAMETER name="t2" default="t"/>
  <PARAMETER name="mu" default="0"/>
  <PARAMETER name="cm" default="mu+V0+V1+V2"/>
  <BASIS ref="hardcore boson"/>
  <SITETERM site="i">
    <PARAMETER name="mu#" default="cm"/>
    -mu#*n(i)
  </SITETERM>
  <BONDTERM source="i" target="j">
    <PARAMETER name="t#" default="0"/>
    <PARAMETER name="V#" default="0"/>
    -t#*(bdag(i)*b(j)+bdag(j)*b(i)) + V#*n(i)*n(j)
  </BONDTERM>
</HAMILTONIAN>

What I use the parameter t=0, I got the following error message:

Internal error: size of prob table is zero.

How could I solve the problem and simulate the model with t equals to 0?

Many thanks!

Looking forward to your reply.

Yours sincerely,

Jianming