[
https://issues.apache.org/jira/browse/SPARK-32907?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17212092#comment-17212092
]
Apache Spark commented on SPARK-32907:
--------------------------------------
User 'zhengruifeng' has created a pull request for this issue:
https://github.com/apache/spark/pull/30009
> adaptively blockify instances
> -----------------------------
>
> Key: SPARK-32907
> URL: https://issues.apache.org/jira/browse/SPARK-32907
> Project: Spark
> Issue Type: Sub-task
> Components: ML
> Affects Versions: 3.1.0
> Reporter: zhengruifeng
> Priority: Major
> Attachments: blockify_svc_perf_20201010.xlsx
>
>
> According to the performance test in
> https://issues.apache.org/jira/browse/SPARK-31783, the performance gain is
> mainly related to the nnz of block.
> So it is reasonable to control the size of block.
>
> I had some offline discuss with [~weichenxu123], then we think following
> changes are worthy：
> 1, infer an appropriate blockSize (MB) based on numFeatures and nnz by
> default;
> 2, impls should use a relative small memory footprint when processing one
> block, and should not use a large pre-allocated buffer, so we need to revert
> gmm;
> 3, use new blockify strategy in LinearSVC/LoR/LiR/AFT;
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
