You are viewing a plain text version of this content. The canonical link for it is here.

Posted to jira@arrow.apache.org by "Neal Richardson (Jira)" <ji...@apache.org> on 2021/09/08 21:02:00 UTC

[jira] [Commented] (ARROW-13940) [R] Multi-threading with Arrow engine queries?

    [ https://issues.apache.org/jira/browse/ARROW-13940?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17412186#comment-17412186 ] 

Neal Richardson commented on ARROW-13940:
-----------------------------------------

Existing queries would have been affected by the change because all dplyr queries now use the ExecPlan interface instead of the Scanner. ExecPlans are run here: https://github.com/apache/arrow/blob/e9251b015184753c53f9a26f32e8f89a8d21a760/r/src/compute-exec.cpp#L58-L85

Maybe there's some multithreading setup that isn't being configured? [~bkietz] [~westonpace]

> [R] Multi-threading with Arrow engine queries?
> ----------------------------------------------
>
>                 Key: ARROW-13940
>                 URL: https://issues.apache.org/jira/browse/ARROW-13940
>             Project: Apache Arrow
>          Issue Type: Bug
>          Components: R
>            Reporter: Jonathan Keane
>            Priority: Major
>
> Since ARROW-13740 was merged, we're seeing dataset queries take close to 2x longer on conbench https://conbench.ursa.dev/benchmarks/e54ae362090b4a868bee929d45936400/
> I'm also seeing only one core utilized when running queries locally as well.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)