triqs.gfs.gf_factories.gf_factories_dlr_imfreq.find_w_max

triqs.gfs.gf_factories.gf_factories_dlr_imfreq.find_w_max()

Dispatched C++ function(s).

[1] (g: Gf[MeshImFreq, 0],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[2] (g: BlockGf[MeshImFreq, 0],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[3] (g: Gf[MeshImFreq, 1],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[4] (g: BlockGf[MeshImFreq, 1],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[5] (g: Gf[MeshImFreq, 2],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[6] (g: BlockGf[MeshImFreq, 2],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[7] (g: Gf[MeshImFreq, 3],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[8] (g: BlockGf[MeshImFreq, 3],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[9] (g: Gf[MeshImFreq, 4],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

[10] (g: BlockGf[MeshImFreq, 4],
     eps: float = 1e-10,
     symmetrize: bool = True,
     w_max_init: float = 1,
     w_max_max: float = 200)
  -> float

Find the smallest DLR energy cutoff that reproduces a Matsubara Green’s function within a target accuracy.

Applies to every overload. Starting from w_max_init, the cutoff is grown by a factor of 1.5 per iteration. For each candidate cutoff, the function builds a DLR Matsubara mesh from (w_max, eps, symmetrize), samples g on that DLR mesh, reconstructs the Green’s function on the original Matsubara mesh via the DLR coefficients, and measures the worst-case round-trip error across Matsubara indices and target indices. The smallest w_max not exceeding w_max_max whose round-trip error is below eps is returned.

Candidate cutoffs whose DLR frequency range exceeds the Matsubara range of the input are skipped. For a block Green’s function, the reported per-iteration error is the worst case across all blocks. An exception is raised if w_max_init exceeds w_max_max, or if no candidate cutoff up to w_max_max achieves an error below eps.

Parameters:

gGf[MeshImFreq, 0]

Input Green’s function on a Matsubara frequency mesh.

epsfloat

Target DLR accuracy and round-trip error tolerance.

symmetrizebool

If true, use particle-hole symmetric DLR meshes.

w_max_initfloat

Initial value of the DLR cutoff to try (must not exceed w_max_max).

w_max_maxfloat

Maximum DLR cutoff to try before giving up.

Returns:

float

The smallest w_max not exceeding w_max_max for which the round-trip error is below eps.