andygrove opened a new issue, #1966: URL: https://github.com/apache/arrow-rs/issues/1966
**Is your feature request related to a problem or challenge? Please describe what you are trying to do.** I would like to be able to leverage GPUs for accelerated compute on Arrow arrays. **Describe the solution you'd like** I have a proof-of-concept that I plan on cleaning up and contributing in the next 1-2 weeks and it provides the following functionality: - unsafe ffi wrapper around the [cuDF](https://github.com/rapidsai/cudf) library that allows raw data, validity, and offset buffers to be transferred between host memory and GPU memory and then allows GPU kernels to be invoked - safe arrow-rs API that supports the transfer of arrow-rs arrays between host memory and GPU memory and invoking GPU kernels on them One of the goals here is that the ffi library is not tied to the arrow-rs implementation so that it can be used by other Arrow implementations (such as `arrow2`) and also by applications that are not using Arrow. **Describe alternatives you've considered** None **Additional context** None -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
