systemml-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Deron Eriksson (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SYSTEMML-1568) NULL condition not check for Spark version in MLContext
Date Mon, 01 May 2017 19:55:04 GMT

    [ https://issues.apache.org/jira/browse/SYSTEMML-1568?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15991424#comment-15991424
] 

Deron Eriksson commented on SYSTEMML-1568:
------------------------------------------

[~niketanpansare]

It looks like this was introduced by:
commit 738ddfc17ca56d7c366bbea83f584e9f10cd8d44
[SYSTEMML-692] Separated Caffe2dml into a separate jar

After this commit, I see:
{code}
scala> val ml = new MLContext(spark)
17/05/01 12:43:21 WARN MLContext: Apache Spark null or above is recommended for SystemML null

Welcome to Apache SystemML!
{code}

With the previous commit, the null message wasn't there:
commit 39a37ae40c84d934236a7748252b196fb0a8b08d
[SYSTEMML-769] Adding support for native BLAS for Linux

{code}
scala> val ml = new MLContext(spark)

Welcome to Apache SystemML!
{code}

Note that there are other issues with commit 738ddfc17ca56d7c366bbea83f584e9f10cd8d44. For
example, wink and antlr runtime are not packaged in the main jar after this commit.

cc [~mwdusenb@us.ibm.com]


> NULL condition not check for Spark version in MLContext
> -------------------------------------------------------
>
>                 Key: SYSTEMML-1568
>                 URL: https://issues.apache.org/jira/browse/SYSTEMML-1568
>             Project: SystemML
>          Issue Type: Bug
>            Reporter: Niketan Pansare
>            Assignee: Deron Eriksson
>            Priority: Minor
>
> I see following warning after starting pyspark shell:
> {code}
> 17/04/30 14:05:25 WARN MLContext: Apache Spark null or above is recommended for SystemML
null
> Welcome to Apache SystemML!
> {code}
> To reproduce the warning, please use Spark 2.1:
> {code}
> # checkout current master
> mvn package -P distribution
> pip install target/systemml-1.0.0-incubating-SNAPSHOT-python.tgz
> pyspark
> >> run simple script with python mlcontext
> {code} 
> [~deron] Can you please take a look at it ?



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message