File size: 7,180 Bytes
c902ee7 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 | #!/usr/bin/env python3
"""Uniform-grid cost diagnostic for the common receiver solvers."""
import os
for key in ('OPENBLAS_NUM_THREADS', 'OMP_NUM_THREADS', 'MKL_NUM_THREADS', 'NUMEXPR_NUM_THREADS'):
os.environ[key] = '2'
import argparse
from datetime import datetime, timezone
import hashlib
import importlib.util
import json
from pathlib import Path
import platform
import sys
import time
import numpy as np
import scipy
from threadpoolctl import threadpool_info, threadpool_limits
HERE = Path(__file__).resolve().parent
ROOT = HERE.parent
spec = importlib.util.spec_from_file_location('receiver_cost_solver', ROOT/'total_error_cost'/'run.py')
SOLVER = importlib.util.module_from_spec(spec)
sys.modules[spec.name] = SOLVER
spec.loader.exec_module(SOLVER)
def digest(path):
return hashlib.sha256(path.read_bytes()).hexdigest()
def hashes():
paths = [HERE/'config.json', HERE/'run.py', ROOT/'total_error_cost'/'run.py',
ROOT/'total_error_cost'/'config.json', ROOT/'accuracy_cost'/'run.py',
ROOT/'accuracy_cost'/'config.json', ROOT/'acoustic_core.py',
ROOT/'propagation'/'chebyshev_reference.py']
return {str(path.relative_to(ROOT)): digest(path) for path in paths}
def dump(path, value):
temporary = path.with_suffix(path.suffix+'.tmp')
temporary.write_text(json.dumps(value, indent=2, allow_nan=False)+'\n')
temporary.replace(path)
def candidates(cfg):
for material in cfg['materials']:
for n in cfg['grids']:
for cfl in cfg['explicit_cfl_values']:
yield {'material': material, 'grid': n, 'method': 'full_stormer_verlet',
'requested_cfl': cfl, 'modes_per_patch': None, 'dimension': n*n}
yield {'material': material, 'grid': n, 'method': 'craig_bampton',
'requested_cfl': None, 'modes_per_patch': cfg['modes_per_patch'],
'dimension': n+2*cfg['modes_per_patch']}
def stem(candidate):
suffix = ('_cfl'+str(candidate['requested_cfl']).replace('.', 'p')
if candidate['method']=='full_stormer_verlet' else '_m'+str(candidate['modes_per_patch']))
return f"{candidate['material']}_n{candidate['grid']}_{candidate['method']}"+suffix
def utc(epoch):
return datetime.fromtimestamp(epoch, timezone.utc).isoformat()
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--dry-run', action='store_true')
args = parser.parse_args()
cfg = json.loads((HERE/'config.json').read_text())
cases = list(candidates(cfg))
dependencies = hashes()
if args.dry_run:
print(json.dumps({'settings':len(cases), 'repetitions':cfg['repetitions'],
'dependency_sha256':dependencies, 'candidates':cases}, indent=2))
return
requested = set(cfg['cpu_affinity'])
if not requested.issubset(os.sched_getaffinity(0)):
raise RuntimeError('Requested CPUs unavailable; do not substitute timing hardware')
os.sched_setaffinity(0, requested)
out = HERE/'results'/'candidates'
out.mkdir(parents=True, exist_ok=True)
for path in out.glob('*_timing.json'):
previous = json.loads(path.read_text())
if previous['dependency_sha256'] != dependencies or previous['cpu_affinity'] != cfg['cpu_affinity']:
raise RuntimeError('Stored timings have different dependencies or affinity')
started = time.time()
environment = {'platform':platform.platform(), 'python':platform.python_version(),
'numpy':np.__version__, 'scipy':scipy.__version__, 'cpu_affinity':sorted(os.sched_getaffinity(0)),
'threads':cfg['threads'], 'dependency_sha256':dependencies, 'started_unix':started,
'started_utc':utc(started), 'settings':len(cases), 'repetitions':cfg['repetitions'],
'cpu_information':Path('/proc/cpuinfo').read_text().split('\n\n')[0],
'thread_environment':{key:os.environ[key] for key in ('OPENBLAS_NUM_THREADS','OMP_NUM_THREADS','MKL_NUM_THREADS')}}
times = np.linspace(0, cfg['time_end'], cfg['samples'])
with threadpool_limits(limits=cfg['threads']):
environment['threadpools'] = threadpool_info()
dump(HERE/'results'/'environment.json', environment)
warm_times = np.linspace(0, .2, 17)
SOLVER.explicit_receiver(8, 'rotated_stripes', .25, warm_times)
SOLVER.explicit_receiver(8, 'rotated_stripes', .125, warm_times)
SOLVER.reduced_receiver(8, 'rotated_stripes', 4, 'craig_bampton', warm_times)
for index, candidate in enumerate(cases):
name = stem(candidate)
target = out/f'{name}_timing.json'
if target.exists():
continue
repetitions, intervals, diagnostics, traces = [], [], [], []
case_start = time.time()
for repetition in range(cfg['repetitions']):
rep_start = time.time()
if candidate['method']=='full_stormer_verlet':
raw, stages, diagnostic = SOLVER.explicit_receiver(candidate['grid'], candidate['material'], candidate['requested_cfl'], times)
else:
raw, stages, diagnostic = SOLVER.reduced_receiver(candidate['grid'], candidate['material'], candidate['modes_per_patch'], candidate['method'], times)
rep_end = time.time()
repetitions.append(stages)
intervals.append({'start_unix':rep_start, 'end_unix':rep_end,
'start_utc':utc(rep_start), 'end_utc':utc(rep_end)})
diagnostics.append(diagnostic)
traces.append(raw['receiver_pressure'].copy())
case_end = time.time()
raw['receiver_pressure_repetitions'] = np.stack(traces)
temporary = out/f'{name}.npz.tmp'
with temporary.open('wb') as stream:
np.savez_compressed(stream, **raw)
raw_path = out/f'{name}.npz'
temporary.replace(raw_path)
record = {**candidate, 'repetitions':repetitions, 'repetition_intervals':intervals,
'median':SOLVER.summarize(repetitions), 'diagnostics':diagnostics,
'cpu_affinity':cfg['cpu_affinity'], 'dependency_sha256':dependencies,
'timed_interval_unix':[case_start,case_end],
'maximum_repeated_trace_difference':max(float(np.linalg.norm(y-traces[-1])/np.linalg.norm(traces[-1])) for y in traces),
'raw_sha256':digest(raw_path), 'receiver_shape':list(traces[-1].shape),
'full_pressure_dimension':candidate['grid']**2}
dump(target, record)
print(json.dumps({'completed':index+1, 'settings':len(cases), 'candidate':name,
'median_seconds':record['median']['total_seconds']}), flush=True)
if hashes()!=dependencies:
raise RuntimeError('A runtime dependency changed during timing')
environment.update(finished_unix=time.time(), finished_utc=utc(time.time()),
completed_settings=len(list(out.glob('*_timing.json'))))
dump(HERE/'results'/'environment.json', environment)
if __name__=='__main__':
main()
|