[
https://issues.apache.org/jira/browse/ARROW-4212?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17661236#comment-17661236
]
Rok Mihevc commented on ARROW-4212:
-----------------------------------
This issue has been migrated to [issue
#20792|https://github.com/apache/arrow/issues/20792] on GitHub. Please see the
[migration documentation|https://github.com/apache/arrow/issues/14542] for
further details.
> [Python] [CUDA] Creating a CUDA buffer from Numba device array should be
> easier
> -------------------------------------------------------------------------------
>
> Key: ARROW-4212
> URL: https://issues.apache.org/jira/browse/ARROW-4212
> Project: Apache Arrow
> Issue Type: Improvement
> Components: GPU, Python
> Affects Versions: 0.11.1
> Reporter: Antoine Pitrou
> Assignee: Pearu Peterson
> Priority: Major
> Labels: pull-request-available
> Fix For: 0.13.0
>
> Time Spent: 20h 10m
> Remaining Estimate: 0h
>
> Currently, to create a CUDA buffer from a Numba device array, you have to
> write:
> {code:python}
> cuda.CudaBuffer.from_numba(device_arr.gpu_data)
> {code}
> It would be easier if you could just write:
> {code}
> cuda.CudaBuffer.from_numba(device_arr)
> {code}
> (ideally, any object exposing the [CUDA Array
> Interface|https://numba.pydata.org/numba-doc/latest/cuda/cuda_array_interface.html]
> would work)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
