[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16144027#comment-16144027 ] Apache Spark commented on SPARK-21688: -- User 'VinceShieh' has created a pull request for this issue:

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16125407#comment-16125407 ] Sean Owen commented on SPARK-21688: --- Solving part of the problem is better than none at all. I do agree

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16125353#comment-16125353 ] Vincent commented on SPARK-21688: - sorry for late reply. Yes, It's simple and easy to check the env

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121679#comment-16121679 ] Sean Owen commented on SPARK-21688: --- Not a good solution? How about just checking the env variables?

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121624#comment-16121624 ] Vincent commented on SPARK-21688: - Okay. Yes, true. It can still run without issue but we are just

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121599#comment-16121599 ] Sean Owen commented on SPARK-21688: --- I mean best case in that MKL might be a little different from

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121589#comment-16121589 ] Vincent commented on SPARK-21688: - [~srowen] Thanks for your comments. I think if user decides to use

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121299#comment-16121299 ] Sean Owen commented on SPARK-21688: --- Understood, though it potentially impacts the benchmarks. You have

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121284#comment-16121284 ] Peng Meng commented on SPARK-21688: --- MKL is just an example of native BLAS, if user has Openblas,

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121270#comment-16121270 ] Sean Owen commented on SPARK-21688: --- I guess my concern is that this slows things down unless people do

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121254#comment-16121254 ] Vincent commented on SPARK-21688: - and if native blas is left with default multi-threading setting, it

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121236#comment-16121236 ] Vincent commented on SPARK-21688: - upload a data we collected before, uni-test on ddot, we can see for

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121225#comment-16121225 ] Sean Owen commented on SPARK-21688: --- I see, so you're saying use BLAS for level 1 ops. Do we know

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121209#comment-16121209 ] Vincent commented on SPARK-21688: - currently, there are certain places in ML/MLLib, such as in mllib/SVM,

[jira] [Commented] (SPARK-21688) performance improvement in mllib SVM with native BLAS

[ https://issues.apache.org/jira/browse/SPARK-21688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121163#comment-16121163 ] Sean Owen commented on SPARK-21688: --- Of course native BLAS is typically faster where it is used so you