Skip to content

Add test for real FFT of misaligned input arrays #3087

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
antonwolfy merged 3 commits into from
Oct 6, 2026
Merged

Add test for real FFT of misaligned input arrays #3087

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
24cdfdf
Add test for real FFT of misaligned input arrays
antonwolfy Oct 5, 2026
aa4c74c
Scale float32 atol with DC magnitude in test_rfft_input_alignment
antonwolfy Oct 5, 2026
1f5d978
Merge branch 'master' into add-rfft-input-alignment-test
antonwolfy Oct 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
42 changes: 41 additions & 1 deletion dpnp/tests/third_party/cupy/fft_tests/test_fft.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,13 +1,14 @@
from __future__ import annotations

import functools
import math
import warnings

import numpy as np
import pytest

import dpnp as cupy
from dpnp.tests.helper import has_support_aspect64
from dpnp.tests.helper import get_float_dtypes, has_support_aspect64

# from cupy.fft import config
# from cupy.fft._fft import (
Expand Down Expand Up @@ -1288,6 +1289,45 @@ def test_irfftn(self, xp, dtype, order, enable_nd):
return xp.fft.irfftn(a, s=self.s, axes=self.axes, norm=self.norm)


@pytest.mark.parametrize("dtype", get_float_dtypes())
@pytest.mark.parametrize("offset", [0, 1])
@pytest.mark.parametrize("ndim", [1, 3])
@pytest.mark.parametrize("layout", ["view", "strided", "pointer"])
def test_rfft_input_alignment(
dtype: type[np.float32 | np.float64], offset: int, ndim: int, layout: str
) -> None:
shape: tuple[int, ...] = (35,) * ndim
backing: cupy.ndarray
x: cupy.ndarray
if layout == "view":
backing = cupy.ones(shape=(2,) + shape, dtype=dtype)
x = backing[offset]
elif layout == "strided":
backing = cupy.ones(shape=shape[:-1] + (2 * shape[-1],), dtype=dtype)
x = backing[..., offset::2]
else:
backing = cupy.ones(shape=math.prod(shape) + 1, dtype=dtype)
x = cupy.ndarray(shape, dtype=dtype, buffer=backing, offset=offset)
assert x.data.ptr % (2 * x.itemsize) == offset * x.itemsize

out: cupy.ndarray = cupy.fft.rfft(x) if ndim == 1 else cupy.fft.rfftn(x)
expected: np.ndarray = np.fft.rfftn(np.ones(shape=shape, dtype=dtype))
# oneMKL f32 err in zero bins scales with DC value (prod(shape))
testing.assert_allclose(
actual=out,
desired=expected,
rtol=1e-5 if dtype is np.float32 else 1e-12,
atol=(
np.finfo(dtype).eps * math.prod(shape)
if dtype is np.float32
else 1e-9
),
)
testing.assert_array_equal(
actual=backing, desired=np.ones(shape=backing.shape, dtype=dtype)
)


# Only those tests in which a legit plan can be obtained are kept
@testing.with_requires("numpy>=2.0")
@pytest.mark.usefixtures("skip_forward_backward")
Expand Down
Loading