[
https://issues.apache.org/jira/browse/SPARK-21152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16060603#comment-16060603
]
Yanbo Liang commented on SPARK-21152:
-------------------------------------
[~sethah] This is an interesting topic, thanks for working on it. Could you
show the performance comparison with respect to the size and type of data?
AFAIK, the most common use case for MLlib LR is training on {{low dimensional
dense/sparse or high dimensional sparse}} data. If blocked gradient update can
get significant performance improvement for these cases, I think it's worth the
investment. Thanks.
> Use level 3 BLAS operations in LogisticAggregator
> -------------------------------------------------
>
> Key: SPARK-21152
> URL: https://issues.apache.org/jira/browse/SPARK-21152
> Project: Spark
> Issue Type: Sub-task
> Components: ML
> Affects Versions: 2.1.1
> Reporter: Seth Hendrickson
>
> In logistic regression gradient update, we currently compute by each
> individual row. If we blocked the rows together, we can do a blocked gradient
> update which leverages the BLAS GEMM operation.
> On high dimensional dense datasets, I've observed ~10x speedups. The problem
> here, though, is that it likely won't improve the sparse case so we need to
> keep both implementations around, and this blocked algorithm will require
> caching a new dataset of type:
> {code}
> BlockInstance(label: Vector, weight: Vector, features: Matrix)
> {code}
> We have avoided caching anything beside the original dataset passed to train
> in the past because it adds memory overhead if the user has cached this
> original dataset for other reasons. Here, I'd like to discuss whether we
> think this patch would be worth the investment, given that it only improves a
> subset of the use cases.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
