[
https://issues.apache.org/jira/browse/ARROW-1951?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16306499#comment-16306499
]
ASF GitHub Bot commented on ARROW-1951:
---------------------------------------
robertnishihara commented on a change in pull request #1451: ARROW-1951:
[Python] Add memcopy threads argument to PlasmaClient put.
URL: https://github.com/apache/arrow/pull/1451#discussion_r159098807
##########
File path: python/pyarrow/plasma.pyx
##########
@@ -370,7 +370,7 @@ cdef class PlasmaClient:
object_buffers[i].metadata_size))
return result
- def put(self, object value, ObjectID object_id=None,
+ def put(self, object value, ObjectID object_id=None, int memcopy_threads=6,
Review comment:
Perhaps as an upper bound. In some cases I think that will be too high,
e.g., on a 64 core machine I did a few experiments and didn't see benefits past
20 threads. Something like `min(NUM_CPUS, 16)` might make sense.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
> Add memcopy_threads to serialization context
> --------------------------------------------
>
> Key: ARROW-1951
> URL: https://issues.apache.org/jira/browse/ARROW-1951
> Project: Apache Arrow
> Issue Type: Improvement
> Reporter: Robert Nishihara
> Assignee: Robert Nishihara
> Priority: Minor
> Labels: pull-request-available
> Fix For: 0.9.0
>
>
> Right now, when calling {{put}} with a plasma client, we set
> {{memcopy_threads}} to 4. We should expose this so it can be changed through
> Python.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
