Added more options for checking PMLs

这个提交包含在:
Craig Warren
2023-08-31 14:47:47 +01:00
父节点 596dc6e330
当前提交 631d9a8414
共有 5 个文件被更改,包括 248 次插入135 次删除

查看文件

@@ -0,0 +1,93 @@
# Copyright (C) 2015-2023: The University of Edinburgh, United Kingdom
# Authors: Craig Warren, Antonis Giannopoulos, and John Hartley
#
# This file is part of gprMax.
#
# gprMax 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 of the License, or
# (at your option) any later version.
#
# gprMax 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 gprMax. If not, see <http://www.gnu.org/licenses/>.
import logging
from pathlib import Path
import h5py
import numpy as np
logger = logging.getLogger(__name__)
def diff_output_files(filename1, filename2):
"""Calculates differences between two output files.
Args:
filename1: string of filename (including path) of output file 1.
filename2: string of filename (including path) of output file 2.
Returns:
time: numpy array containing time.
datadiffs: numpy array containing power (dB) of differences.
"""
file1 = h5py.File(Path(filename1), "r")
file2 = h5py.File(Path(filename2), "r")
# Path to receivers in files
path = "rxs/rx1/"
# Get available field output component names
outputs1 = list(file1[path].keys())
outputs2 = list(file2[path].keys())
if outputs1 != outputs2:
logger.exception("Field output components are not the same in each file")
raise ValueError
# Check that type of float used to store fields matches
floattype1 = file1[path + outputs1[0]].dtype
floattype2 = file2[path + outputs2[0]].dtype
if floattype1 != floattype2:
logger.warning(
f"Type of floating point number in test model ({file1[path + outputs1[0]].dtype}) "
f"does not match type in reference solution ({file2[path + outputs2[0]].dtype})\n"
)
# Arrays for storing time
time1 = np.zeros((file1.attrs["Iterations"]), dtype=floattype1)
time1 = np.linspace(0, (file1.attrs["Iterations"] - 1), num=file1.attrs["Iterations"])
time2 = np.zeros((file2.attrs["Iterations"]), dtype=floattype2)
time2 = np.linspace(0, (file2.attrs["Iterations"] - 1), num=file2.attrs["Iterations"])
# Arrays for storing field data
data1 = np.zeros((file1.attrs["Iterations"], len(outputs1)), dtype=floattype1)
data2 = np.zeros((file2.attrs["Iterations"], len(outputs2)), dtype=floattype2)
for ID, name in enumerate(outputs1):
data1[:, ID] = file1[path + str(name)][:]
data2[:, ID] = file2[path + str(name)][:]
if np.any(np.isnan(data1[:, ID])) or np.any(np.isnan(data2[:, ID])):
logger.exception("Data contains NaNs")
raise ValueError
file1.close()
file2.close()
# Diffs
datadiffs = np.zeros(data1.shape, dtype=np.float64)
for i in range(len(outputs2)):
maxi = np.amax(np.abs(data1[:, i]))
datadiffs[:, i] = np.divide(
np.abs(data2[:, i] - data1[:, i]), maxi, out=np.zeros_like(data1[:, i]), where=maxi != 0
) # Replace any division by zero with zero
# Calculate power (ignore warning from taking a log of any zero values)
with np.errstate(divide="ignore"):
datadiffs[:, i] = 20 * np.log10(datadiffs[:, i])
# Replace any NaNs or Infs from zero division
datadiffs[:, i][np.invert(np.isfinite(datadiffs[:, i]))] = 0
return time1, datadiffs

查看文件

@@ -1,135 +0,0 @@
# Copyright (C) 2015-2023: The University of Edinburgh, United Kingdom
# Authors: Craig Warren, Antonis Giannopoulos, and John Hartley
#
# This file is part of gprMax.
#
# gprMax 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 of the License, or
# (at your option) any later version.
#
# gprMax 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 gprMax. If not, see <http://www.gnu.org/licenses/>.
import itertools
import logging
from operator import add
from pathlib import Path
import h5py
import matplotlib.pyplot as plt
import numpy as np
logger = logging.getLogger(__name__)
# Create/setup plot figure
# colors = ['#E60D30', '#5CB7C6', '#A21797', '#A3B347'] # Plot colours from http://tools.medialab.sciences-po.fr/iwanthue/index.php
# colorIDs = ["#62a85b", "#9967c7", "#b3943f", "#6095cd", "#cb5c42", "#c95889"]
colorIDs = ["#79c72e", "#5774ff", "#ff7c2c", "#4b4e80", "#d7004e", "#007545", "#ff83ec"]
# colorIDs = ["#ba0044", "#b2d334", "#470055", "#185300", "#ff96b1", "#3e2700", "#0162a9", "#fdb786"]
colors = itertools.cycle(colorIDs)
# for i in range(2):
# next(colors)
lines = itertools.cycle(("--", ":", "-.", "-"))
markers = ["o", "d", "^", "s", "*"]
parts = Path(__file__).parts
path = "rxs/rx1/"
basename = "pml_3D_pec_plate"
PMLIDs = ["CFS-PML", "HORIPML-1", "HORIPML-2", "MRIPML-1", "MRIPML-2"]
maxerrors = []
testmodels = ["pml_3D_pec_plate_" + s for s in PMLIDs]
fig, ax = plt.subplots(
subplot_kw=dict(xlabel="Iterations", ylabel="Error [dB]"), figsize=(20, 10), facecolor="w", edgecolor="w"
)
for x, model in enumerate(testmodels):
# Open output file and read iterations
fileref = h5py.File(Path(*parts[:-1], basename, basename + "_ref.h5"), "r")
filetest = h5py.File(Path(*parts[:-1], basename, basename + str(x + 1) + ".h5"), "r")
# Get available field output component names
outputsref = list(fileref[path].keys())
outputstest = list(filetest[path].keys())
if outputsref != outputstest:
logger.exception("Field output components do not match reference solution")
raise ValueError
# Check that type of float used to store fields matches
if filetest[path + outputstest[0]].dtype != fileref[path + outputsref[0]].dtype:
logger.warning(
f"Type of floating point number in test model ({filetest[path + outputstest[0]].dtype}) "
f"does not match type in reference solution ({fileref[path + outputsref[0]].dtype})\n"
)
floattyperef = fileref[path + outputsref[0]].dtype
floattypetest = filetest[path + outputstest[0]].dtype
# logger.info(f'Data type: {floattypetest}')
# Arrays for storing time
# timeref = np.zeros((fileref.attrs['Iterations']), dtype=floattyperef)
# timeref = np.linspace(0, (fileref.attrs['Iterations'] - 1) * fileref.attrs['dt'], num=fileref.attrs['Iterations']) / 1e-9
# timetest = np.zeros((filetest.attrs['Iterations']), dtype=floattypetest)
# timetest = np.linspace(0, (filetest.attrs['Iterations'] - 1) * filetest.attrs['dt'], num=filetest.attrs['Iterations']) / 1e-9
timeref = np.zeros((fileref.attrs["Iterations"]), dtype=floattyperef)
timeref = np.linspace(0, (fileref.attrs["Iterations"] - 1), num=fileref.attrs["Iterations"])
timetest = np.zeros((filetest.attrs["Iterations"]), dtype=floattypetest)
timetest = np.linspace(0, (filetest.attrs["Iterations"] - 1), num=filetest.attrs["Iterations"])
# Arrays for storing field data
dataref = np.zeros((fileref.attrs["Iterations"], len(outputsref)), dtype=floattyperef)
datatest = np.zeros((filetest.attrs["Iterations"], len(outputstest)), dtype=floattypetest)
for ID, name in enumerate(outputsref):
dataref[:, ID] = fileref[path + str(name)][:]
datatest[:, ID] = filetest[path + str(name)][:]
if np.any(np.isnan(datatest[:, ID])):
logger.exception("Test data contains NaNs")
raise ValueError
fileref.close()
filetest.close()
# Diffs
datadiffs = np.zeros(datatest.shape, dtype=np.float64)
for i in range(len(outputstest)):
maxi = np.amax(np.abs(dataref[:, i]))
datadiffs[:, i] = np.divide(
np.abs(datatest[:, i] - dataref[:, i]), maxi, out=np.zeros_like(dataref[:, i]), where=maxi != 0
) # Replace any division by zero with zero
# Calculate power (ignore warning from taking a log of any zero values)
with np.errstate(divide="ignore"):
datadiffs[:, i] = 20 * np.log10(datadiffs[:, i])
# Replace any NaNs or Infs from zero division
datadiffs[:, i][np.invert(np.isfinite(datadiffs[:, i]))] = 0
# Print maximum error value
start = 210
maxerrors.append(f": {np.amax(datadiffs[start::, 1]):.1f} [dB]")
logger.info(f"{model}: Max. error {maxerrors[x]}")
# Plot diffs (select column to choose field component, 0-Ex, 1-Ey etc..)
ax.plot(timeref[start::], datadiffs[start::, 1], color=next(colors), lw=2, ls=next(lines), label=model)
ax.set_xticks(np.arange(0, 2200, step=100))
ax.set_xlim([0, 2100])
ax.set_yticks(np.arange(-160, 0, step=20))
ax.set_ylim([-160, -20])
ax.set_axisbelow(True)
ax.grid(color=(0.75, 0.75, 0.75), linestyle="dashed")
mylegend = list(map(add, PMLIDs, maxerrors))
legend = ax.legend(mylegend, loc=1, fontsize=14)
frame = legend.get_frame()
frame.set_edgecolor("white")
frame.set_alpha(0)
plt.show()
# Save a PDF/PNG of the figure
# fig.savefig(basepath + '.pdf', dpi=None, format='pdf', bbox_inches='tight', pad_inches=0.1)
# fig.savefig(savename + '.png', dpi=150, format='png', bbox_inches='tight', pad_inches=0.1)

查看文件

@@ -0,0 +1,80 @@
# Copyright (C) 2015-2023: The University of Edinburgh, United Kingdom
# Authors: Craig Warren, Antonis Giannopoulos, and John Hartley
#
# This file is part of gprMax.
#
# gprMax 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 of the License, or
# (at your option) any later version.
#
# gprMax 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 gprMax. If not, see <http://www.gnu.org/licenses/>.
import itertools
import logging
from operator import add
from pathlib import Path
import matplotlib.pyplot as plt
import numpy as np
from testing.diff_output_files import diff_output_files
logger = logging.getLogger(__name__)
# Create/setup plot figure
# colors = ['#E60D30', '#5CB7C6', '#A21797', '#A3B347'] # Plot colours from http://tools.medialab.sciences-po.fr/iwanthue/index.php
# colorIDs = ["#62a85b", "#9967c7", "#b3943f", "#6095cd", "#cb5c42", "#c95889"]
colorIDs = ["#79c72e", "#5774ff", "#ff7c2c", "#4b4e80", "#d7004e", "#007545", "#ff83ec"]
# colorIDs = ["#ba0044", "#b2d334", "#470055", "#185300", "#ff96b1", "#3e2700", "#0162a9", "#fdb786"]
colors = itertools.cycle(colorIDs)
# for i in range(2):
# next(colors)
lines = itertools.cycle(("--", ":", "-.", "-"))
markers = ["o", "d", "^", "s", "*"]
fn = Path(__file__)
basename = "pml_3D_pec_plate"
PMLIDs = ["CFS-PML", "HORIPML-1", "HORIPML-2", "MRIPML-1", "MRIPML-2"]
maxerrors = []
testmodels = [basename + "_" + s for s in PMLIDs]
fig, ax = plt.subplots(
subplot_kw=dict(xlabel="Iterations", ylabel="Error [dB]"), figsize=(20, 10), facecolor="w", edgecolor="w"
)
for x, model in enumerate(testmodels):
time, datadiffs = diff_output_files(fn.parent.joinpath(basename + "_ref.h5"),
fn.parent.joinpath(basename + str(x + 1) + ".h5"))
# Print maximum error value
start = 210
maxerrors.append(f": {np.amax(datadiffs[start::, 1]):.1f} [dB]")
logger.info(f"{model}: Max. error {maxerrors[x]}")
# Plot diffs (select column to choose field component, 0-Ex, 1-Ey etc..)
ax.plot(time[start::], datadiffs[start::, 1], color=next(colors), lw=2, ls=next(lines), label=model)
ax.set_xticks(np.arange(0, 2200, step=100))
ax.set_xlim([0, 2100])
ax.set_yticks(np.arange(-160, 0, step=20))
ax.set_ylim([-160, -20])
ax.set_axisbelow(True)
ax.grid(color=(0.75, 0.75, 0.75), linestyle="dashed")
mylegend = list(map(add, PMLIDs, maxerrors))
legend = ax.legend(mylegend, loc=1, fontsize=14)
frame = legend.get_frame()
frame.set_edgecolor("white")
frame.set_alpha(0)
plt.show()
# Save a PDF/PNG of the figure
# fig.savefig(basepath + '.pdf', dpi=None, format='pdf', bbox_inches='tight', pad_inches=0.1)
# fig.savefig(savename + '.png', dpi=150, format='png', bbox_inches='tight', pad_inches=0.1)

查看文件

@@ -0,0 +1,75 @@
# Copyright (C) 2015-2023: The University of Edinburgh, United Kingdom
# Authors: Craig Warren, Antonis Giannopoulos, and John Hartley
#
# This file is part of gprMax.
#
# gprMax 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 of the License, or
# (at your option) any later version.
#
# gprMax 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 gprMax. If not, see <http://www.gnu.org/licenses/>.
import itertools
import logging
from operator import add
from pathlib import Path
import matplotlib.pyplot as plt
import numpy as np
from testing.diff_output_files import diff_output_files
logger = logging.getLogger(__name__)
# Create/setup plot figure
# Plot colours from http://tools.medialab.sciences-po.fr/iwanthue/index.php
colorIDs = ["#79c72e", "#5774ff", "#ff7c2c", "#4b4e80", "#d7004e", "#007545", "#ff83ec"]
colors = itertools.cycle(colorIDs)
lines = itertools.cycle(("--", ":", "-.", "-"))
markers = ["o", "d", "^", "s", "*"]
fn = Path(__file__)
basename = "pml_basic"
PMLIDs = ["off", "x0", "y0", "z0", "xmax", "ymax", "zmax"]
maxerrors = []
for x in range(len(PMLIDs)):
file1 = fn.parent.joinpath(basename + str(x + 1) + "_CPU.h5")
file2 = fn.parent.joinpath(basename + str(x + 1) + "_GPU.h5")
time, datadiffs = diff_output_files(file1, file2)
# Print maximum error value
start = 0
maxerrors.append(f": {np.amax(datadiffs[start::, 1]):.1f} [dB]")
logger.info(f"{file1.name} - {file2.name}: Max. error {maxerrors[x]}")
fig, ax = plt.subplots(
subplot_kw=dict(xlabel="Iterations", ylabel="Error [dB]"), figsize=(20, 10), facecolor="w", edgecolor="w")
# Plot diffs (select column to choose field component, 0-Ex, 1-Ey etc..)
ax.plot(time[start::], datadiffs[start::, 1], color=next(colors), lw=2, ls=next(lines), label=f"{file1.name} - {file2.name}")
ax.set_xticks(np.arange(0, 2200, step=100))
ax.set_xlim([0, 2100])
ax.set_yticks(np.arange(-160, 0, step=20))
ax.set_ylim([-160, -20])
ax.set_axisbelow(True)
ax.grid(color=(0.75, 0.75, 0.75), linestyle="dashed")
mylegend = list(map(add, PMLIDs, maxerrors))
legend = ax.legend(mylegend, loc=1, fontsize=14)
frame = legend.get_frame()
frame.set_edgecolor("white")
frame.set_alpha(0)
plt.show()
# Save a PDF/PNG of the figure
# fig.savefig(basepath + '.pdf', dpi=None, format='pdf', bbox_inches='tight', pad_inches=0.1)
# fig.savefig(savename + '.png', dpi=150, format='png', bbox_inches='tight', pad_inches=0.1)