# Error using if numba cuda nvidia bindings are used

**URL:** <https://numba.discourse.group/t/error-using-if-numba-cuda-nvidia-bindings-are-used/1867>\
**Category:** Support: What is this error message?\
**Created:** [March 30, 2023, 2:03pm UTC](https://numba.discourse.group/t/error-using-if-numba-cuda-nvidia-bindings-are-used/1867 "2023-03-30T14:03:27Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![biosckon](https://yyz2.discourse-cdn.com/free1/user_avatar/numba.discourse.group/biosckon/32/775_2.png) [@biosckon](https://numba.discourse.group/u/biosckon)\
**Post date:** [March 30, 2023, 2:03pm UTC](https://numba.discourse.group/t/error-using-if-numba-cuda-nvidia-bindings-are-used/1867/1 "2023-03-30T14:03:27Z")

</div>

Follow up to post [CUDA - OpenGL interop](https://numba.discourse.group/t/cuda-opengl-interop/1865/7)

Setting **`NUMBA_CUDA_USE_NVIDIA_BINDING=1`** before running the code  
below results in error.

Un-setting the var does not produce error.

```python
import ctypes
import numpy as np
from cuda import cudart
from cuda.cudart import cudaGraphicsRegisterFlags as GLFlags
import numba.cuda as nbcuda
import moderngl as mgl

###################################################
# wrap the raw pointer into DeviceNDArray
def gl_NDArray(ptr, shape, dtype=np.float32, strides=None, order="C"):
    shape, strides, dtype = nbcuda.api.prepare_shape_strides_dtype(
        shape, strides, dtype, order
    )
    datasize = nbcuda.driver.memory_size_from_info(shape, strides, dtype.itemsize)

    ctx = nbcuda.current_context()
    c_ptr = ctypes.c_uint64(ptr)
    mem = nbcuda.driver.MemoryPointer(ctx, c_ptr, datasize)
    return nbcuda.cudadrv.devicearray.DeviceNDArray(shape, strides, dtype, gpu_data=mem)

ctx = mgl.create_context(standalone=True)

# vertices for triangle with Red Green and Blue colors
vertices = np.array(
    # X Y x y x y
    [-0.6, -0.6, 1.0, 0.0, 0.0, 0.6, -0.6, 0.0, 1.0, 0.0, 0.0, 0.6, 0.0, 0.0, 1.0],
    dtype="f4",
)

# init VBO
vbo = ctx.buffer(vertices)

# Get CUDA pointer to VBO
err, gl_resource = cudart.cudaGraphicsGLRegisterBuffer(
    vbo.glo, GLFlags.cudaGraphicsRegisterFlagsNone 
)
(err,) = cudart.cudaGraphicsMapResources(1, gl_resource, 0)

# get raw pointer to device array
(err, dev_ptr, dev_ptr_size) = cudart.cudaGraphicsResourceGetMappedPointer(gl_resource)
# wrap into DeviceNDArray
vbo_arr = gl_NDArray(dev_ptr, vertices.size)

# fails here
a = vbo_arr.copy_to_host()
print(a)

```

Below error happens on if `NUMBA_CUDA_USE_NVIDIA_BINDING=1`.

```auto
Traceback (most recent call last):
  File "C:\Users\k1m19\OneDrive\Desktop\CUDA_RAPIDS\debug_code1.py", line 48, in <module>
    a = vbo_arr.copy_to_host()
  File "D:\PY\mamba\envs\cuda_learn_1\lib\site-packages\numba\cuda\cudadrv\devices.py", line 232, in _require_cuda_context
    return fn(*args, **kws)
  File "D:\PY\mamba\envs\cuda_learn_1\lib\site-packages\numba\cuda\cudadrv\devicearray.py", line 277, in copy_to_host
    _driver.device_to_host(hostary, self, self.alloc_size,
  File "D:\PY\mamba\envs\cuda_learn_1\lib\site-packages\numba\cuda\cudadrv\driver.py", line 3230, in device_to_host
    fn(host_pointer(dst), device_pointer(src), size, *varargs)
  File "D:\PY\mamba\envs\cuda_learn_1\lib\site-packages\numba\cuda\cudadrv\driver.py", line 361, in safe_cuda_api_call
    return self._check_cuda_python_error(fname, libfn(*args))
  File "cuda\cuda.pyx", line 18103, in cuda.cuda.cuMemcpyDtoH
  File "cuda\cuda.pyx", line 3785, in cuda.cuda.CUdeviceptr. __cinit__
TypeError: an integer is required

```