TeNPy Forum

First of all, let me clarify, also for those finding this question later: The system purely 1D / we ignore the y direction, right? I.e. when we talk about periodic boundary conditions, we don't talk about wrapping things on a cylinder or torus or plane, but about a chain (open boundary conditions, O...

When you need chi=256 for L=8, you apparently have local hilbert space dimension of 4 - is it a spin-full fermion site? Do you have any charge conservation/symmetries in general? DMRG should usually be able to converge L=32/64 pretty easily. One issue you can have though are close to exact degenerac...

1. Yes - as noted in the comment at the top of the xxz_chain.py implementation. 2. You can always go back in the git history and find it there with git log --all -- examples/xxz_groundstate.py which brings up d1dccf29ee50899451ea95931c5f4216efae8e56 .' 3. For finite, critical 1D systems, the famouse...

Whit error bar you mean "what values do I consider degenerate"? Don't choose that value too small based on what the norm error is: the error in the entanglement spectrum depends drastically on how high the value is. If you plot the full spectrum up to the chi_max, the highest values in the...

Disclaimer: You already said that MPS are not a good idea for this kind of model... That said, I can point you to you misunderstanding: the add_coupling method is for adding a bunch of terms at once, all of the form \sum_i J_{i, i+dx} A_i B_{i+dx} . You want to add a single term A_i B_j , which is w...

There's no simple, generic rule how large/small dt has to be to ensure convergence, I'm somewhat alarmed by the "periodic boundary conditions". By default, in TeNPy this would just give you one (or ~Ly, if you have a 2D torus) very long-range couplings between the ends of the MPS. Both TDV...

This is ill-defined for infinite systems: The overlap \braket{s | t} for any two iMPS which do not represent exactly the same state goes to 0 in the the thermodynamic limit (or to infinity, if you don't have normalized MPS.). Also, you generally don't want to write down "cat state" superpo...

You should be able to get to significantly larger system sizes even on desktops. Typical run times range from several hours up to a couple of days (up to one or two weeks) on typical desktops / single-node cluster machines, depending on what's available to you. The usual geometry of choice are long,...

I'm afraid we can't help much there. How did you install Python? For windows, in my experience the anaconda suite works best, for which you should be able to directly install the corresponding physics-tenpy conda package rather than install it from source. Can you import numpy directly from within y...

The picture you drew above is also not an inner product with the remaining open legs - do you think about closing those as well? That would mean that T is the translation operator with *periodic* boundary conditions. It's not too hard to contract all of this on the tensor level, though. I started wr...

Indeed, Markus is right that singlets are only one example of entangled states. But they're only entangled on pairs of sites, so for more than 2 sites they're "product states" again. Really, you need to clarify which entangled state you want - there are plenty. For example, you could also ...

There is a good reason for this, since it's exponentially expensive with the number of sites L in the segment, both in memory \mathcal{O}(d^{2L}) and CPU time \mathcal{O}(d^{3L}) ! Each additional site quadruples your computational cost. If you're reaching that limit, it might be better to use the e...

Yes, for the DM interaction, you need to write out the cross product explicitly and add the Dx/Dy/Dz terms separately. Note that you got the signs in the cross product wrong, c.f. the code below. Again, I'm not sure if I understood correctly - it seems you want the D to depend on which pairs of NN y...

The XX model has charge conservation of total Sz. In the zero-particle sector, you can map it exactly to a single-particle fermion hopping, and the ground state is the W state. However, this is not true in the sector of higher filling when you have peridoic boundary conditions: for the coupling goin...

You can write this analytically as a bond-dimension 2 MPS. You can see that you need bond dimension 2, since you always have either exactly one or no 1 on the left/right of any of the two cuts (between the first two sites, or the second two sites). The Schmidt decomposition between the last two site...

I'm not sure why you say this. The from_infiniteT method should already work with other sites, including FermionSite or SpinHalfFermionSite. It's straightforward to adjust the purification.py example to use another model like e.g. the tenpy.models.fermions.FermionChain and measure N instead of Sz. T...

You can still use the tenpy.networks.site.set_common_charges (the multi_sites_combine_charges is deprecated) to do that. If you specify new_charges='independent' , it does exactly what you're asking for: it conserves the total charge on the f1 sites separate from the one of the f2 site. I'm not sure...

Not sure what your question is - you already have a loop over the coupling_axis , that you didn't include in your sample code. Just read out different model_params and explicitly use those instead of the plain J1/J2 for the various terms, e.g. def init_terms(self, model_params): Jxx = model_params.g...

Indeed, by adding the extra spin site on the center, you get a next-nearest neighbor hopping, so you can't use a NearestNeighborModel anymore. There's two ways to proceed: 1) Drop the inheritance from the NearestNeighborModel to not generate the H_bond. This means you can't use TEBD anymore, but the...

very naively, you can multiply out all the products such that you're left with a sum of terms, many of which are long-range. Less naively, you'd try to rewrite it in the nearest-neighbor version O_k = \prod_{\langle i, k\rangle} (1-n_i n_k ) B_k mentioned above, and multiply that out to get sums of...

Ho did you get O(t)? Currently, I don't think there is TEBD code to Heisenberg-Evolve an operator - did you write that yourself? I presume you have O in the form of an MPO then? For a TeNPy MPO, it's simply expectation_value , e.g. E = H.expectation_value(psi) , or for finite systems MPOEnvironment(...

At a glimplse, this looks horrible for DMRG, because each term is non-local, giving you an immense MPO bond dimension if you implement it naively. Do you have only those terms, i.e. is H=\left(\prod_{\langle i, j \rangle} (1-n_i n_j) \right) H^0 with a "usual" H^0 ? It looks like you're tr...

That's currently not supported at the moment, as imaginary time evolution is generally much slower than variational methods like DMRG.
Why would you prefer TDVP over DMRG?

The tenpy.models.tj_model.tJModel defines the options cons_N and cons_Sz - just set them as you need them, and make sure your initial state is in the charge sector you want - for the initial product state you wrote, you get average density of 2/4 = 1/2 and spin Sz=0. You can also easily define a cus...

The MPO tensors in W are sparse, and without the 'sort_mpo_legs': True, we try to capture that, resulting in fewer small blocks for tensordots involving a H.get_W(i) rather than bigger blocks (as dictated by charge conservation). For this benchmark comparison, we wanted to compare "apples to ap...