muSpectre/1c871f98ffb7arcpatch-D169
arcpatch-D169 vs master
Commit | Author | Details | Committed | ||||
---|---|---|---|---|---|---|---|
0b724252e1c5 | pastewka | MAINT: Only print from rank 0 | Mar 18 2018 | ||||
0c68249a13f3 | pastewka | MAINT: Added barrier before finalizing MPI. | Mar 18 2018 | ||||
a9f0a70f9921 | pastewka | ENH: Expose cell factories for different FFT engines to Python. | Mar 18 2018 | ||||
d4aa1fdbde07 | pastewka | CMAKE: Define WITH_FFTWMPI, WITH_PFFT and WITH_P3DFFT depending on which… | Mar 18 2018 | ||||
6317516b3a68 | pastewka | ENH: Added ParallelCellFactory to Python bindings. MPI communicator needs to be… | Mar 17 2018 | ||||
df8be21ac1c8 | pastewka | ENH: Added method that returns size of communicator. | Mar 17 2018 | ||||
cadcc9dc12b4 | pastewka | CMAKE: Removed finding of FFT packages from top level CMake file. | Mar 17 2018 | ||||
7290eee6503e | pastewka | Merge branch 'arcpatch-D169' into feat/mpi | Mar 16 2018 | ||||
ec30870e6e86 | D169 | junge | fixed python tests | Mar 16 2018 |
/
README
µSpectre
Copyright © 2018 Till Junge
µSpectre is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3, or (at your option) any later version.
µSpectre is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with GNU Emacs; see the file COPYING. If not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
Building µSpectre
µSpectre is a CMake project that uses C++14. It depends on the Boost unit test framework for testing and uses uses python3 as secondary API. You will need a modern C++ compiler (µSpectre was tested with gcc-6 , gcc-7, clang-4 and clang5) and CMake version 3.7.0 or higher.
Compilation
$ git clone https://c4science.ch/source/muSpectre.git $ cd build $ cmake -DCMAKE_BUILD_TYPE=Release .. $ make
µSpectre makes use of expression templates, as a result, production code must be compiled with the CMAKE_BUILD_TYPE flag set to Release in order to get performance (under gcc, non-optimised code is about 50 times slower than the release). Be careful with parallel compilation: compiling µSpectre is quite memory-hungry because of the use of expression templates, a parallel make -j requires currently about 10 GB of RAM under GCC.
Simple usage example
The following is a simple example for using µSpectre through its convenient Python interface
#!/usr/bin/env python3 import numpy as np import pyMuSpectre as µ # setting the geometry resolution = [51, 51] center = np.array([r//2 for r in resolution]) incl = resolution[0]//5 lengths = [7., 5.] formulation = µ.Formulation.small_strain # creating the periodic cell rve = µ.SystemFactory(resolution, lengths, formulation) hard = µ.material.MaterialHooke2d.make( rve, "hard", 10e9, .33) soft = µ.material.MaterialHooke2d.make( rve, "soft", 70e9, .33) # assign a material to each pixel for i, pixel in enumerate(rve): if np.linalg.norm(center - np.array(pixel),2)<incl: hard.add_pixel(pixel) else: soft.add_pixel(pixel) tol = 1e-5 cg_tol = 1e-8 # set macroscopic strain Del0 = np.array([[.0, .0], [0, .03]]) if formulation == µ.Formulation.small_strain: Del0 = .5*(Del0 + Del0.T) maxiter = 401 verbose = 2 solver = SolverCG(rve, cg_tol, maxiter, verbose=False) r = µ.solvers.newton_cg(rve, Del0, solver, tol, verbose) print("nb of {} iterations: {}".format(solver.name(), r.nb_fev))
You can find more examples using both the python and the c++ interface in the bin/ and tests folder.