Can't use cupy.ndarray's __setitem__ when the value is an array-like object that implements __cuda_array_interface__ or __cupy_get_ndarray__ but also implements __eq__. #9089
Description
Description
Can't use cupy.ndarray's __setitem__ when the value is an array-like object that implements __cuda_array_interface__ or __cupy_get_ndarray__ but also implements __eq__.
The reproducer below works with numpy
x = Array(np.array([1, 1, 2]))
y = np.array([1, 2, 3])
y[:] = x
ySo I expect it to work for cupy as well and not raise
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
Cell In[33], line 28
26 x = Array(cp.array([1, 1, 2]))
27 y = cp.array([1, 2, 3])
---> 28 y[:] = x
29 y
File cupy/_core/core.pyx:1641, in cupy._core.core._ndarray_base.__setitem__()
File cupy/_core/_routines_indexing.pyx:51, in cupy._core._routines_indexing._ndarray_setitem()
File cupy/_core/_routines_indexing.pyx:1017, in cupy._core._routines_indexing._scatter_op()
File cupy/_core/core.pyx:789, in cupy._core.core._ndarray_base.fill()
File cupy/_core/core.pyx:1293, in cupy._core.core._ndarray_base.__nonzero__()
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()
To Reproduce
import numpy as np
import cupy as cp
class Array:
def __init__(self, data):
self.data = data
def __array__(self, *args, **kwargs):
return self.data.__array__(*args, **kwargs)
def __cupy_get_ndarray__(self, *args, **kwargs):
return self.data
@property
def __array_interface__(self):
return self.data.__array_interface__
@property
def __cuda_array_interface__(self):
return self.data.__cuda_array_interface__
def __eq__(self, other):
return self.data == other
x = Array(cp.array([1, 1, 2]))
y = cp.array([1, 2, 3])
y[:] = x
yDeleting the __eq__ implementation works fine
import numpy as np
import cupy as cp
class Array:
def __init__(self, data):
self.data = data
def __array__(self, *args, **kwargs):
return self.data.__array__(*args, **kwargs)
def __cupy_get_ndarray__(self, *args, **kwargs):
return self.data
@property
def __array_interface__(self):
return self.data.__array_interface__
@property
def __cuda_array_interface__(self):
return self.data.__cuda_array_interface__
x = Array(cp.array([1, 1, 2]))
y = cp.array([1, 2, 3])
y[:] = x
yand returns array([1, 1, 2])
Installation
Wheel (pip install cupy-***)
Environment
OS : Linux-6.11.0-21-generic-x86_64-with-glibc2.39
Python Version : 3.12.9
CuPy Version : 13.4.1
CuPy Platform : NVIDIA CUDA
NumPy Version : 2.2.4
SciPy Version : 1.15.2
Cython Build Version : 3.0.12
Cython Runtime Version : None
CUDA Root : /usr/local/cuda
nvcc PATH : /usr/local/cuda/bin/nvcc
CUDA Build Version : 12080
CUDA Driver Version : 12080
CUDA Runtime Version : 12080 (linked to CuPy) / 12080 (locally installed)
CUDA Extra Include Dirs : []
cuBLAS Version : 120804
cuFFT Version : 11303
cuRAND Version : 10309
cuSOLVER Version : (11, 7, 3)
cuSPARSE Version : 12508
NVRTC Version : (12, 8)
Thrust Version : 200800
CUB Build Version : 200800
Jitify Build Version : <unknown>
cuDNN Build Version : None
cuDNN Version : None
NCCL Build Version : None
NCCL Runtime Version : None
cuTENSOR Version : 20200
cuSPARSELt Build Version : None
Device 0 Name : NVIDIA GeForce RTX 4080 SUPER
Device 0 Compute Capability : 89
Device 0 PCI Bus ID : 0000:01:00.0
Additional Information
No response