wyj
229f50cd86
Replace the nested spatial global convolution in fast_psf_accumulator_resolve with a reusable double-precision FFTW linear convolution on the CPU PSF backend.
- Add the private src/fast_psf_fftw.{c,h} module: zero-padded R2C/C2R plans, cached kernel spectrum and planar scratch, exact 1/(Pwidth*Pheight) and 1/N^2 normalization, (R,R) crop, and additive HDR output.
- Keep the previous nested loops as fast_psf_accumulator_resolve_spatial_reference for tests/benchmarks only; it is not a runtime fallback.
- Cache the circular row spans on FastPsfAccumulator and report one-time plan, kernel transform, scratch, and per-frame stage timings.
- Require fftw3_omp for CPU builds; HIP and dummy builds do not link FFTW.
- Namespace test/helper binaries by spacetime and build tag, and reject make test / psf-capture for non-CPU backends.
- Add tests/test_fast_psf_fftw.c (FFTW versus spatial), tests/benchmark_fast_psf_fftw.c, an FFTW CLI smoke check, and the 2026-09-25 benchmark record.
182 lines
9.5 KiB
Python
182 lines
9.5 KiB
Python
#!/usr/bin/env python3
|
|
"""Exercise camera defaults/errors and single-frame/movie agreement (CPU PNG builds)."""
|
|
import os
|
|
from pathlib import Path
|
|
import struct
|
|
import subprocess
|
|
import sys
|
|
import tempfile
|
|
import zlib
|
|
|
|
BUILD = Path(sys.argv[1] if len(sys.argv) > 1 else 'build/Release').resolve()
|
|
TESTDIR = Path(sys.argv[2]).resolve() if len(sys.argv) > 2 else BUILD
|
|
ENV = dict(os.environ, OMP_NUM_THREADS='4')
|
|
|
|
|
|
def run(binary, *args, ok=True, env=ENV):
|
|
result = subprocess.run([str(binary), *map(str, args)], env=env,
|
|
capture_output=True, text=True)
|
|
if (result.returncode == 0) != ok:
|
|
raise AssertionError(f'{binary.name} {args}: {result.returncode}\n{result.stderr}')
|
|
return result
|
|
|
|
|
|
def image_payload(path, dimensions=(64, 48), allow_black=False):
|
|
data = path.read_bytes()
|
|
assert data[:8] == b'\x89PNG\r\n\x1a\n'
|
|
offset, compressed = 8, bytearray()
|
|
while offset < len(data):
|
|
count, kind = struct.unpack_from('>I4s', data, offset)
|
|
payload = data[offset + 8:offset + 8 + count]
|
|
if kind == b'IHDR':
|
|
assert struct.unpack_from('>II', payload) == dimensions
|
|
if kind == b'IDAT':
|
|
compressed.extend(payload)
|
|
offset += count + 12
|
|
raw = zlib.decompress(compressed)
|
|
assert raw, f'missing image data: {path}'
|
|
assert allow_black or any(raw), f'empty image: {path}'
|
|
return raw
|
|
|
|
|
|
def map_vertices(path):
|
|
data = path.read_bytes()
|
|
assert data[:8] == b'GRLENS\x01\x00'
|
|
assert struct.unpack_from('<Q', data, 32)[0] == 1
|
|
vertices, triangles = struct.unpack_from('<QQ', data, 64)
|
|
offset = 80
|
|
values = []
|
|
for _ in range(vertices):
|
|
values.append(struct.unpack_from('<9dI', data, offset))
|
|
offset += 76
|
|
return values, data[offset:offset + triangles * 28]
|
|
|
|
|
|
with tempfile.TemporaryDirectory(prefix='gr-camera-cli-') as directory:
|
|
tmp = Path(directory)
|
|
for backend in ('minkowski', 'schwarzschild'):
|
|
binary = BUILD / f'{backend}_sky'
|
|
help_text = run(binary, '--help').stdout
|
|
for option in ('--observer-position', '--observer-velocity', '--camera-roll-deg'):
|
|
assert option in help_text
|
|
assert '--observer-inward-speed' not in help_text
|
|
common = ['--catalog', 'assets/sky_grid_5deg.csv', '--width', 64,
|
|
'--height', 48, '--fov-deg', 80, '--exposure', 0.1,
|
|
'--coarse-cell-pixels', 8, '--refine-max-level', 0, '--psf-relative-tail', 1e-4]
|
|
def render(name, *options):
|
|
path = tmp / f'{backend}_{name}.png'
|
|
run(binary, *common, '--output', path, *options)
|
|
return image_payload(path)
|
|
|
|
# CPU fast-mode CLI smoke test: the FFTW resolve must run and report its
|
|
# one-time setup line.
|
|
if backend == 'minkowski':
|
|
fast_path = tmp / 'minkowski_fast.png'
|
|
fast = run(binary, *common, '--fast-mode', '--fast-supersample', 2,
|
|
'--output', fast_path)
|
|
assert 'Fast FFTW:' in fast.stderr, fast.stderr
|
|
assert image_payload(fast_path)
|
|
|
|
# Equivalent independently specified and inferred camera geometry.
|
|
inferred = render('position', '--observer-position', -30, 0, 0)
|
|
explicit = render('explicit', '--observer-position', -30, 0, 0,
|
|
'--look-ra-deg', 0, '--look-dec-deg', 0)
|
|
angled = render('angle', '--look-ra-deg', 0, '--look-dec-deg', 0)
|
|
assert inferred == explicit == angled
|
|
pole = render('pole', '--observer-position', 0, 0, 30)
|
|
assert pole == render('pole_explicit', '--observer-position', 0, 0, 30,
|
|
'--look-ra-deg', 0, '--look-dec-deg', -90)
|
|
default = render('default')
|
|
pos = (0, 0, 0) if backend == 'minkowski' else (0, 0, 30)
|
|
assert default == render('default_explicit', '--observer-position', *pos,
|
|
'--look-ra-deg', 90, '--look-dec-deg', -90)
|
|
assert render('radius', '--observer-radius', 40) == render(
|
|
'radius_explicit', '--observer-radius', 40, '--look-ra-deg', 90,
|
|
'--look-dec-deg', -90)
|
|
for partial, value, ra, dec in [('--look-ra-deg', 37, 37, -90),
|
|
('--look-dec-deg', -23, 90, -23)]:
|
|
assert render('partial', partial, value) == render(
|
|
'complete', '--look-ra-deg', ra, '--look-dec-deg', dec)
|
|
errors = [
|
|
(['--observer-position', 1, 2], None),
|
|
(['--observer-position', 1, 2, 'nan'], None),
|
|
(['--observer-velocity', 0, 0, 'inf'], None),
|
|
(['--look-ra-deg', 'nan'], None),
|
|
(['--look-dec-deg', 'inf'], None),
|
|
(['--observer-radius', 'nan'], None),
|
|
(['--observer-radius', 0], None),
|
|
(['--camera-roll-deg', 'nan'], None),
|
|
(['--observer-position', 0, 0, 0], 'Cannot infer'),
|
|
(['--observer-position', 3, 4, 5, '--observer-radius', 30], 'mutually exclusive'),
|
|
(['--observer-velocity', 10, 0, 0], 'not timelike'),
|
|
(['--observer-inward-speed', 0], None),
|
|
(['--observer-track', 'missing.csv', '--observer-velocity', 0, 0, 0], 'cannot be combined'),
|
|
(['--frames-dir', tmp, '--look-ra-deg', 0], 'cannot be combined'),
|
|
(['--lens-map-input', 'missing.grlens', '--camera-roll-deg', 0], 'cannot be combined'),
|
|
]
|
|
if backend == 'schwarzschild':
|
|
errors += [(['--observer-position', 1.5, 0, 0, '--observer-velocity', -0.5, 0, 0], 'capture cutoff'),
|
|
(['--observer-position', 1.75, 0, 0], 'not timelike')]
|
|
render('inside', '--observer-position', 1.75, 0, 0,
|
|
'--observer-velocity', -0.5, 0, 0, '--look-ra-deg', 0, '--look-dec-deg', 0)
|
|
for options, message in errors:
|
|
missing_catalog = tmp / 'should_not_be_created.csv'
|
|
result = run(binary, '--catalog', missing_catalog, *options, ok=False)
|
|
if message:
|
|
assert message in result.stderr, result.stderr
|
|
assert not missing_catalog.exists(), result.stderr
|
|
assert 'PSF cache ready' not in result.stderr
|
|
track = tmp / f'{backend}.csv'
|
|
run(TESTDIR / f'test_observer_{backend}', track)
|
|
single_map, movie_map = tmp / 'single.grlens', tmp / 'movie.grlens'
|
|
single = render('moving', '--observer-position', 3, -4, 5,
|
|
'--observer-velocity', 0.2, -0.1, 0.3,
|
|
'--look-ra-deg', 37, '--look-dec-deg', -23,
|
|
'--camera-roll-deg', 19, '--lens-map-output', single_map)
|
|
run(binary, *common, '--observer-track', track, '--frames-dir', tmp,
|
|
'--frames-prefix', backend, '--duration', 0, '--fps', 1,
|
|
'--lens-map-output', movie_map)
|
|
movie = image_payload(tmp / f'{backend}_000000.png')
|
|
assert single == movie, f'{backend}: single/movie PNG mismatch'
|
|
a, ta = map_vertices(single_map)
|
|
b, tb = map_vertices(movie_map)
|
|
assert len(a) == len(b) and ta == tb
|
|
max_error = 0
|
|
for x, y in zip(a, b):
|
|
assert x[-1] == y[-1], 'ray classification mismatch'
|
|
max_error = max(max_error, *(abs(v - w) for v, w in zip(x[:-1], y[:-1])))
|
|
assert max_error < 1e-9, max_error
|
|
# A map import must still work without evaluating a camera/metric.
|
|
assert single == render('import', '--lens-map-input', single_map)
|
|
if backend == 'schwarzschild':
|
|
# Two inward-looking free-fall samples at r=6.2696 and r=3.1593.
|
|
# At 16:9 the latter frame finishes in generation 0, while the
|
|
# former still needs refinement. Finishing the empty batch used
|
|
# to abort the whole movie in generation 1.
|
|
mixed = Path(__file__).parent / 'fixtures/schwarzschild_mixed_refinement.csv'
|
|
parallel_map = tmp / 'mixed-parallel.grlens'
|
|
result = run(binary, *common, '--height', 36, '--fov-deg', 60,
|
|
'--refine-max-level', 3, '--observer-track', mixed,
|
|
'--movie-track-samples', '--frames-dir', tmp,
|
|
'--frames-prefix', 'mixed', '--verbose',
|
|
'--lens-map-output', parallel_map)
|
|
assert 'Ray trace generation 1: frame 0 added' in result.stderr
|
|
assert 'Ray trace generation 0: frame 1 added' not in result.stderr
|
|
for frame in range(2):
|
|
image_payload(tmp / f'mixed_{frame:06d}.png',
|
|
dimensions=(64, 36), allow_black=True)
|
|
# Thread scheduling must preserve endpoints, frame/sample IDs and
|
|
# the resulting adaptive mesh across the entire slab sweep.
|
|
for threads in (1, 16):
|
|
comparison_map = tmp / f'mixed-{threads}-threads.grlens'
|
|
run(binary, *common, '--height', 36, '--fov-deg', 60,
|
|
'--refine-max-level', 3, '--observer-track', mixed,
|
|
'--movie-track-samples', '--frames-dir', tmp,
|
|
'--frames-prefix', f'mixed-{threads}-threads',
|
|
'--lens-map-output', comparison_map,
|
|
env=dict(ENV, OMP_NUM_THREADS=str(threads)))
|
|
assert parallel_map.read_bytes() == comparison_map.read_bytes(), \
|
|
f'movie lens map changed with {threads} threads'
|
|
print('schwarzschild: movie lens map identical with 1, 4 and 16 threads', flush=True)
|
|
print(f'{backend}: CLI checks passed; single/movie PNG identical, map max error {max_error:.3g}', flush=True)
|