Coverage for pySDC/projects/TOMS/AllenCahn_contracting_circle.py: 95%
184 statements
« prev ^ index » next coverage.py v7.6.7, created at 2024-11-16 14:51 +0000
1import os
3import dill
4import matplotlib.ticker as ticker
5import numpy as np
7import pySDC.helpers.plot_helper as plt_helper
8from pySDC.helpers.stats_helper import get_sorted
10from pySDC.implementations.controller_classes.controller_nonMPI import controller_nonMPI
11from pySDC.implementations.problem_classes.AllenCahn_2D_FD import (
12 allencahn_fullyimplicit,
13 allencahn_semiimplicit,
14 allencahn_semiimplicit_v2,
15 allencahn_multiimplicit,
16 allencahn_multiimplicit_v2,
17)
18from pySDC.implementations.sweeper_classes.generic_implicit import generic_implicit
19from pySDC.implementations.sweeper_classes.imex_1st_order import imex_1st_order
20from pySDC.implementations.sweeper_classes.multi_implicit import multi_implicit
21from pySDC.projects.TOMS.AllenCahn_monitor import monitor
24# http://www.personal.psu.edu/qud2/Res/Pre/dz09sisc.pdf
27def setup_parameters():
28 """
29 Helper routine to fill in all relevant parameters
31 Note that this file will be used for all versions of SDC, containing more than necessary for each individual run
33 Returns:
34 description (dict)
35 controller_params (dict)
36 """
38 # initialize level parameters
39 level_params = dict()
40 level_params['restol'] = 1e-08
41 level_params['dt'] = 1e-03
42 level_params['nsweeps'] = [1]
44 # initialize sweeper parameters
45 sweeper_params = dict()
46 sweeper_params['quad_type'] = 'RADAU-RIGHT'
47 sweeper_params['num_nodes'] = [3]
48 sweeper_params['Q1'] = ['LU']
49 sweeper_params['Q2'] = ['LU']
50 sweeper_params['QI'] = ['LU']
51 sweeper_params['QE'] = ['EE']
52 sweeper_params['initial_guess'] = 'zero'
54 # This comes as read-in for the problem class
55 problem_params = dict()
56 problem_params['nu'] = 2
57 problem_params['nvars'] = [(128, 128)]
58 problem_params['eps'] = [0.04]
59 problem_params['newton_maxiter'] = 100
60 problem_params['newton_tol'] = 1e-09
61 problem_params['lin_tol'] = 1e-10
62 problem_params['lin_maxiter'] = 100
63 problem_params['radius'] = 0.25
65 # initialize step parameters
66 step_params = dict()
67 step_params['maxiter'] = 50
69 # initialize controller parameters
70 controller_params = dict()
71 controller_params['logger_level'] = 30
72 controller_params['hook_class'] = monitor
74 # fill description dictionary for easy step instantiation
75 description = dict()
76 description['problem_class'] = None # pass problem class
77 description['problem_params'] = problem_params # pass problem parameters
78 description['sweeper_class'] = None # pass sweeper (see part B)
79 description['sweeper_params'] = sweeper_params # pass sweeper parameters
80 description['level_params'] = level_params # pass level parameters
81 description['step_params'] = step_params # pass step parameters
83 return description, controller_params
86def run_SDC_variant(variant=None, inexact=False):
87 """
88 Routine to run particular SDC variant
90 Args:
91 variant (str): string describing the variant
92 inexact (bool): flag to use inexact nonlinear solve (or nor)
94 Returns:
95 timing (float)
96 niter (float)
97 """
99 # load (incomplete) default parameters
100 description, controller_params = setup_parameters()
102 # add stuff based on variant
103 if variant == 'fully-implicit':
104 description['problem_class'] = allencahn_fullyimplicit
105 description['sweeper_class'] = generic_implicit
106 if inexact:
107 description['problem_params']['newton_maxiter'] = 1
108 elif variant == 'semi-implicit':
109 description['problem_class'] = allencahn_semiimplicit
110 description['sweeper_class'] = imex_1st_order
111 if inexact:
112 description['problem_params']['lin_maxiter'] = 10
113 elif variant == 'semi-implicit_v2':
114 description['problem_class'] = allencahn_semiimplicit_v2
115 description['sweeper_class'] = imex_1st_order
116 if inexact:
117 description['problem_params']['newton_maxiter'] = 1
118 elif variant == 'multi-implicit':
119 description['problem_class'] = allencahn_multiimplicit
120 description['sweeper_class'] = multi_implicit
121 if inexact:
122 description['problem_params']['newton_maxiter'] = 1
123 description['problem_params']['lin_maxiter'] = 10
124 elif variant == 'multi-implicit_v2':
125 description['problem_class'] = allencahn_multiimplicit_v2
126 description['sweeper_class'] = multi_implicit
127 if inexact:
128 description['problem_params']['newton_maxiter'] = 1
129 else:
130 raise NotImplementedError('Wrong variant specified, got %s' % variant)
132 if inexact:
133 out = 'Working on inexact %s variant...' % variant
134 else:
135 out = 'Working on exact %s variant...' % variant
136 print(out)
138 # setup parameters "in time"
139 t0 = 0
140 Tend = 0.032
142 # instantiate controller
143 controller = controller_nonMPI(num_procs=1, controller_params=controller_params, description=description)
145 # get initial values on finest level
146 P = controller.MS[0].levels[0].prob
147 uinit = P.u_exact(t0)
149 # call main function to get things done...
150 uend, stats = controller.run(u0=uinit, t0=t0, Tend=Tend)
152 # filter statistics by variant (number of iterations)
153 iter_counts = get_sorted(stats, type='niter', sortby='time')
155 # compute and print statistics
156 niters = np.array([item[1] for item in iter_counts])
157 out = ' Mean number of iterations: %4.2f' % np.mean(niters)
158 print(out)
159 out = ' Range of values for number of iterations: %2i ' % np.ptp(niters)
160 print(out)
161 out = ' Position of max/min number of iterations: %2i -- %2i' % (int(np.argmax(niters)), int(np.argmin(niters)))
162 print(out)
163 out = ' Std and var for number of iterations: %4.2f -- %4.2f' % (float(np.std(niters)), float(np.var(niters)))
164 print(out)
166 print(' Iteration count (nonlinear/linear): %i / %i' % (P.newton_itercount, P.lin_itercount))
167 print(
168 ' Mean Iteration count per call: %4.2f / %4.2f'
169 % (P.newton_itercount / max(P.newton_ncalls, 1), P.lin_itercount / max(P.lin_ncalls, 1))
170 )
172 timing = get_sorted(stats, type='timing_run', sortby='time')
174 print('Time to solution: %6.4f sec.' % timing[0][1])
175 print()
177 return stats
180def show_results(fname, cwd=''):
181 """
182 Plotting routine
184 Args:
185 fname (str): file name to read in and name plots
186 cwd (str): current working directory
187 """
189 file = open(cwd + fname + '.pkl', 'rb')
190 results = dill.load(file)
191 file.close()
193 # plt_helper.mpl.style.use('classic')
194 plt_helper.setup_mpl()
196 # set up plot for timings
197 fig, ax1 = plt_helper.newfig(textwidth=238.96, scale=1.5, ratio=0.4)
199 timings = {}
200 niters = {}
201 for key, item in results.items():
202 timings[key] = get_sorted(item, type='timing_run', sortby='time')[0][1]
203 iter_counts = get_sorted(item, type='niter', sortby='time')
204 niters[key] = np.mean(np.array([item[1] for item in iter_counts]))
206 xcoords = list(range(len(timings)))
207 sorted_timings = sorted([(key, timings[key]) for key in timings], reverse=True, key=lambda tup: tup[1])
208 sorted_niters = [(k, niters[k]) for k in [key[0] for key in sorted_timings]]
209 heights_timings = [item[1] for item in sorted_timings]
210 heights_niters = [item[1] for item in sorted_niters]
211 keys = [(item[0][1] + ' ' + item[0][0]).replace('-', '\n').replace('_v2', ' mod.') for item in sorted_timings]
213 ax1.bar(xcoords, heights_timings, align='edge', width=-0.3, label='timings (left axis)')
214 ax1.set_ylabel('time (sec)')
216 ax2 = ax1.twinx()
217 ax2.bar(xcoords, heights_niters, color='lightcoral', align='edge', width=0.3, label='iterations (right axis)')
218 ax2.set_ylabel('mean number of iterations')
220 ax1.set_xticks(xcoords)
221 ax1.set_xticklabels(keys, rotation=90, ha='center')
223 # ask matplotlib for the plotted objects and their labels
224 lines, labels = ax1.get_legend_handles_labels()
225 lines2, labels2 = ax2.get_legend_handles_labels()
226 ax2.legend(lines + lines2, labels + labels2, loc=0)
228 # save plot, beautify
229 f = fname + '_timings'
230 plt_helper.savefig(f)
232 assert os.path.isfile(f + '.pdf'), 'ERROR: plotting did not create PDF file'
233 # assert os.path.isfile(f + '.pgf'), 'ERROR: plotting did not create PGF file'
234 assert os.path.isfile(f + '.png'), 'ERROR: plotting did not create PNG file'
236 # set up plot for radii
237 fig, ax = plt_helper.newfig(textwidth=238.96, scale=1.0)
239 exact_radii = []
240 for key, item in results.items():
241 computed_radii = get_sorted(item, type='computed_radius', sortby='time')
243 xcoords = [item0[0] for item0 in computed_radii]
244 radii = [item0[1] for item0 in computed_radii]
245 if key[0] + ' ' + key[1] == 'fully-implicit exact':
246 ax.plot(xcoords, radii, label=(key[0] + ' ' + key[1]).replace('_v2', ' mod.'))
248 exact_radii = get_sorted(item, type='exact_radius', sortby='time')
250 diff = np.array([abs(item0[1] - item1[1]) for item0, item1 in zip(exact_radii, computed_radii)])
251 max_pos = int(np.argmax(diff))
252 assert max(diff) < 0.07, 'ERROR: computed radius is too far away from exact radius, got %s' % max(diff)
253 assert 0.028 < computed_radii[max_pos][0] < 0.03, (
254 'ERROR: largest difference is at wrong time, got %s' % computed_radii[max_pos][0]
255 )
257 xcoords = [item[0] for item in exact_radii]
258 radii = [item[1] for item in exact_radii]
259 ax.plot(xcoords, radii, color='k', linestyle='--', linewidth=1, label='exact')
261 ax.yaxis.set_major_formatter(ticker.FormatStrFormatter('%1.2f'))
262 ax.set_ylabel('radius')
263 ax.set_xlabel('time')
264 ax.grid()
265 ax.legend(loc=3)
267 # save plot, beautify
268 f = fname + '_radii'
269 plt_helper.savefig(f)
271 assert os.path.isfile(f + '.pdf'), 'ERROR: plotting did not create PDF file'
272 # assert os.path.isfile(f + '.pgf'), 'ERROR: plotting did not create PGF file'
273 assert os.path.isfile(f + '.png'), 'ERROR: plotting did not create PNG file'
275 # set up plot for interface width
276 fig, ax = plt_helper.newfig(textwidth=238.96, scale=1.0)
278 interface_width = []
279 for key, item in results.items():
280 interface_width = get_sorted(item, type='interface_width', sortby='time')
281 xcoords = [item[0] for item in interface_width]
282 width = [item[1] for item in interface_width]
283 if key[0] + ' ' + key[1] == 'fully-implicit exact':
284 ax.plot(xcoords, width, label=key[0] + ' ' + key[1])
286 xcoords = [item[0] for item in interface_width]
287 init_width = [interface_width[0][1]] * len(xcoords)
288 ax.plot(xcoords, init_width, color='k', linestyle='--', linewidth=1, label='exact')
290 ax.yaxis.set_major_formatter(ticker.FormatStrFormatter('%1.2f'))
291 ax.set_ylabel(r'interface width ($\epsilon$)')
292 ax.set_xlabel('time')
293 ax.grid()
294 ax.legend(loc=3)
296 # save plot, beautify
297 f = fname + '_interface'
298 plt_helper.savefig(f)
300 assert os.path.isfile(f + '.pdf'), 'ERROR: plotting did not create PDF file'
301 # assert os.path.isfile(f + '.pgf'), 'ERROR: plotting did not create PGF file'
302 assert os.path.isfile(f + '.png'), 'ERROR: plotting did not create PNG file'
304 return None
307def main(cwd=''):
308 """
309 Main driver
311 Args:
312 cwd (str): current working directory (need this for testing)
313 """
315 # Loop over variants, exact and inexact solves
316 results = {}
317 for variant in ['multi-implicit', 'semi-implicit', 'fully-implicit', 'semi-implicit_v2', 'multi-implicit_v2']:
318 results[(variant, 'exact')] = run_SDC_variant(variant=variant, inexact=False)
319 results[(variant, 'inexact')] = run_SDC_variant(variant=variant, inexact=True)
321 # dump result
322 fname = 'data/results_SDC_variants_AllenCahn_1E-03'
323 file = open(cwd + fname + '.pkl', 'wb')
324 dill.dump(results, file)
325 file.close()
326 assert os.path.isfile(cwd + fname + '.pkl'), 'ERROR: dill did not create file'
328 # visualize
329 show_results(fname, cwd=cwd)
332if __name__ == "__main__":
333 main()