You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Apache Arrow is an aligned chunked array representation of columns. If my matrix is a dense matrix encoded as a chunked array of float64 numbers is it possible to use faer-rs on top of it in a zero-copy way? I assume further constraints (chunksize has to be a multiple of some batch size) will be required. Or will we have to combine the (relatively large) chunks into a continuous aligned array first before feeding it to faer?
The text was updated successfully, but these errors were encountered:
Apache Arrow is an aligned chunked array representation of columns. If my matrix is a dense matrix encoded as a chunked array of float64 numbers is it possible to use faer-rs on top of it in a zero-copy way? I assume further constraints (chunksize has to be a multiple of some batch size) will be required. Or will we have to combine the (relatively large) chunks into a continuous aligned array first before feeding it to faer?
The text was updated successfully, but these errors were encountered: