solver.vrtx.dat

Introduction

The solver.vrtx.dat file is designed to store the two-particle Green's function $\chi(i\omega_n, i\omega'_n, i\nu_n)$ and vertex function $\mathcal{F}(i\omega_n, i\omega'_n, i\nu_n)$. It will be output by the quantum impurity solvers when they are shut down.

Note

Only the GARDENIA and NARCISSUS components can generate the solver.vrtx.dat file.
The data stored in the solver.twop.dat file are generated using standard algorithm. However, those data in the solver.vrtx.dat file are generated using the improved estimator algorithm. Unfortunately, as for the CAMELLIA, LAVENDER, and MANJUSHAKA components, they don't support the improved estimator algorithm, so they couldn't generate the solver.vrtx.dat file.

Format

The solver.vrtx.dat file contains

\[ \left(\sum\limits^{\text{norbs}}_{n=1} \frac{n(n+1)}{2}\right)\times \text{nbfrq}\]

blocks. Each block is appended by two blank lines. The format of each block is as follows:

column 1: Matsubara frequency point (fermionic type), $\omega_n$, integer, the unit is $\pi/\beta$

column 2: Matsubara frequency point (fermionic type), $\omega'_n$, integer, the unit is $\pi/\beta$

column 3: two-particle Green's function, $\chi(i\omega_n, i\omega'_n, i\nu_n)$, double precision

column 4: bubble function, $\chi_0(i\omega_n, i\omega'_n, i\nu_n)$, double precision

column 5: irreducible part of the two-particle Green's function, $\chi_\text{irr}(i\omega_n, i\omega'_n, i\nu_n)$, double precision

column 6: full vertex function, $\mathcal{F}(i\omega_n, i\omega'_n, i\nu_n)$, double precision

Note

In the solver.vrtx.dat file, we adopt the following orbital sequence: $1\uparrow$, $2\uparrow$, $3\uparrow$, $\cdots$, $1\downarrow$, $2\downarrow$, $3\downarrow$, $\cdots$ In other words, the spin up part is always before the spin down part.

Code

The corresponding Fortran code block for the writing of solver.vrtx.dat file is as follows:

! open data file: solver.vrtx.dat
     open(mytmp, file='solver.vrtx.dat', form='formatted', status='unknown')

! write it
     do m=1,norbs
         do n=1,m
             do k=1,nbfrq
                 write(mytmp,'(a,i6)') '# flvr1:', m
                 write(mytmp,'(a,i6)') '# flvr2:', n
                 write(mytmp,'(a,i6)') '# nbfrq:', k
                 do j=1,nffrq

                     ......

                     do i=1,nffrq

                         ......

                         write(mytmp,'(2i6,8f16.8)') jt, it, chit, chi0, chii, chii/(g1*g2*g3*g4)
                     enddo ! over i={1,nffrq} loop
                 enddo ! over j={1,nffrq} loop
                 write(mytmp,*) ! write empty lines
                 write(mytmp,*)
             enddo ! over k={1,nbfrq} loop
         enddo ! over n={1,m} loop
     enddo ! over m={1,norbs} loop

! close data file
     close(mytmp)

In the HIBISCUS component, we provide a Python module to read the output files of quantum impurity solvers. You can use it to read the solver.vrtx.dat file. See script/u_reader.py for more details.