Skip to content

GitLab

Commit e894cae8 authored by eric pellegrini's avatar eric pellegrini
Browse files

Replaced Lapack dgesdd explicit call for SVD decomposition by numpy.linalg.svd

parent 736c2038
Hide whitespace changes
Inline Side-by-side
MMTK/NormalModes/Core.py
View file @ e894cae8
......@@ -11,18 +11,6 @@ import numpy
from MMTK import ParticleProperties, Visualization
dgesdd = None
try:
from numpy.linalg.lapack_lite import dgesdd
except ImportError:
pass
if dgesdd is None:
try:
# PyLAPACK
from lapack_dge import dgesdd
except ImportError: pass
#
# Class for a single mode
#
......@@ -229,64 +217,21 @@ class NormalModes(object):
ntotal = nexcluded + nmodes
natoms = len(basis[0])
sv = numpy.zeros((min(ntotal, 3*natoms),), numpy.float)
work_size = numpy.zeros((1,), numpy.float)
if nexcluded > 0:
self.basis = numpy.zeros((ntotal, 3*natoms), numpy.float)
for i in range(nexcluded):
self.basis[i] = numpy.ravel(excluded[i].array*self.weights)
min_n_m = min(3*natoms, nexcluded)
vt = numpy.zeros((nexcluded, min_n_m), numpy.float)
iwork = numpy.zeros((8*min_n_m,), int_type)
if 3*natoms >= nexcluded:
result = dgesdd('O', 3*natoms, nexcluded, self.basis,
3*natoms, sv, self.basis, 3*natoms, vt, min_n_m,
work_size, -1, iwork, 0)
work = numpy.zeros((int(work_size[0]),), numpy.float)
result = dgesdd('O', 3*natoms, nexcluded, self.basis,
3*natoms, sv, self.basis, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
else:
u = numpy.zeros((min_n_m, 3*natoms), numpy.float)
result = dgesdd('S', 3*natoms, nexcluded, self.basis,
3*natoms, sv, u, 3*natoms, vt, min_n_m,
work_size, -1, iwork, 0)
work = numpy.zeros((int(work_size[0]),), numpy.float)
result = dgesdd('S', 3*natoms, nexcluded, self.basis,
3*natoms, sv, u, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
self.basis[:min_n_m] = u
if result['info'] != 0:
raise ValueError('Lapack SVD: ' + `result['info']`)
_,sv,self.basis = numpy.linalg(self.basis,False)
svmax = numpy.maximum.reduce(sv)
nexcluded = numpy.add.reduce(numpy.greater(sv, 1.e-10*svmax))
ntotal = nexcluded + nmodes
for i in range(nmodes):
self.basis[i+nexcluded] = numpy.ravel(basis[i].array*self.weights)
min_n_m = min(3*natoms, ntotal)
vt = numpy.zeros((ntotal, min_n_m), numpy.float)
if 3*natoms >= ntotal:
result = dgesdd('O', 3*natoms, ntotal, self.basis, 3*natoms,
sv, self.basis, 3*natoms, vt, min_n_m,
work_size, -1, iwork, 0)
if int(work_size[0]) > work.shape[0]:
work = numpy.zeros((int(work_size[0]),), numpy.float)
result = dgesdd('O', 3*natoms, ntotal, self.basis, 3*natoms,
sv, self.basis, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
else:
u = numpy.zeros((min_n_m, 3*natoms), numpy.float)
result = dgesdd('S', 3*natoms, ntotal, self.basis, 3*natoms,
sv, u, 3*natoms, vt, min_n_m,
work_size, -1, iwork, 0)
if int(work_size[0]) > work.shape[0]:
work = numpy.zeros((int(work_size[0]),), numpy.float)
result = dgesdd('S', 3*natoms, ntotal, self.basis, 3*natoms,
sv, u, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
self.basis[:min_n_m] = u
if result['info'] != 0:
raise ValueError('Lapack SVD: ' + `result['info']`)
_,sv,self.basis = numpy.linalg(self.basis,False)
svmax = numpy.maximum.reduce(sv)
ntotal = numpy.add.reduce(numpy.greater(sv, 1.e-10*svmax))
nmodes = ntotal - nexcluded
......@@ -298,29 +243,9 @@ class NormalModes(object):
self.basis = numpy.array(map(lambda v: v.array, basis))
numpy.multiply(self.basis, self.weights, self.basis)
self.basis.shape = (nmodes, 3*natoms)
min_n_m = min(3*natoms, nmodes)
vt = numpy.zeros((nmodes, min_n_m), numpy.float)
iwork = numpy.zeros((8*min_n_m,), int_type)
if 3*natoms >= nmodes:
result = dgesdd('O', 3*natoms, nmodes, self.basis, 3*natoms,
sv, self.basis, 3*natoms, vt, min_n_m,
work_size, -1, iwork, 0)
work = numpy.zeros((int(work_size[0]),), numpy.float)
result = dgesdd('O', 3*natoms, nmodes, self.basis, 3*natoms,
sv, self.basis, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
else:
u = numpy.zeros((min_n_m, 3*natoms), numpy.float)
result = dgesdd('S', 3*natoms, nmodes, self.basis, 3*natoms,
sv, u, 3*natoms, vt, min_n_m,
work_size, -1, iwork, 0)
work = numpy.zeros((int(work_size[0]),), numpy.float)
result = dgesdd('S', 3*natoms, nmodes, self.basis, 3*natoms,
sv, u, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
self.basis[:min_n_m] = u
if result['info'] != 0:
raise ValueError('Lapack SVD: ' + `result['info']`)
_,sv,self.basis = numpy.linalg(self.basis,False)
svmax = numpy.maximum.reduce(sv)
nmodes = numpy.add.reduce(numpy.greater(sv, 1.e-10*svmax))
ntotal = nmodes
......
MMTK/Subspace.py
View file @ e894cae8
......@@ -23,44 +23,6 @@ from MMTK import Utility, ParticleProperties
from Scientific.Geometry import Vector, ex, ey, ez
import copy
#
# Import LAPACK routines
#
dgesdd = None
try:
from numpy.linalg.lapack_lite import dgesdd, LapackError
except ImportError:
pass
if dgesdd is None:
try:
# PyLAPACK
from lapack_dge import dgesdd
except ImportError: pass
if dgesdd:
n = 1
array = numpy.zeros((n, n), numpy.float)
sv = numpy.zeros((n,), numpy.float)
u = numpy.zeros((n, n), numpy.float)
vt = numpy.zeros((n, n), numpy.float)
work = numpy.zeros((1,), numpy.float)
int_types = [numpy.int, numpy.int8, numpy.int16, numpy.int32]
try:
int_types.append(numpy.int64)
except AttributeError:
pass
for int_type in int_types:
iwork = numpy.zeros((1,), int_type)
try:
dgesdd('S', n, n, array, n, sv, u, n, vt, n, work, -1, iwork, 0)
break
except LapackError:
pass
del n, array, sv, u, vt, work, iwork, int_types
del array_package
#
# A set of particle vectors that define a subspace
#
......@@ -128,36 +90,15 @@ class Subspace(object):
nvectors = shape[0]
natoms = shape[1]
basis.shape = (nvectors, 3*natoms)
sv = numpy.zeros((min(nvectors, 3*natoms),), numpy.float)
min_n_m = min(3*natoms, nvectors)
vt = numpy.zeros((nvectors, min_n_m), numpy.float)
work = numpy.zeros((1,), numpy.float)
iwork = numpy.zeros((8*min_n_m,), int_type)
if 3*natoms >= nvectors:
result = dgesdd('O', 3*natoms, nvectors, basis, 3*natoms,
sv, basis, 3*natoms, vt, min_n_m,
work, -1, iwork, 0)
work = numpy.zeros((int(work[0]),), numpy.float)
result = dgesdd('O', 3*natoms, nvectors, basis, 3*natoms,
sv, basis, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
u = basis
else:
u = numpy.zeros((min_n_m, 3*natoms), numpy.float)
result = dgesdd('S', 3*natoms, nvectors, basis, 3*natoms,
sv, u, 3*natoms, vt, min_n_m,
work, -1, iwork, 0)
work = numpy.zeros((int(work[0]),), numpy.float)
result = dgesdd('S', 3*natoms, nvectors, basis, 3*natoms,
sv, u, 3*natoms, vt, min_n_m,
work, work.shape[0], iwork, 0)
if result['info'] != 0:
raise ValueError('Lapack SVD: ' + `result['info']`)
_,sv,u = numpy.linalg.svd(basis,True)
svmax = numpy.maximum.reduce(sv)
nvectors = numpy.add.reduce(numpy.greater(sv, 1.e-10*svmax))
u = u[:nvectors]
u.shape = (nvectors, natoms, 3)
self._basis = ParticleVectorSet(self.universe, u)
return self._basis
def projectionOf(self, vector):
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment