[jira] [Updated] (MAHOUT-690) LanczosSolver tests take forever. No fun.

["Sean Owen (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/MAHOUT-690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sean Owen updated MAHOUT-690:
-----------------------------

    Attachment: MAHOUT-690.patch

Agree, this is over half the test time and does burn a fair bit of time as I run tests frequently.

One way to speed it up is to shrink the test data size, no? I did so and made it run about 3x faster without causing it to fail, but does this look like it remains a valid test?

The non-CLI test can be shrunk further but then an assertion about the size of error fails.

> LanczosSolver tests take forever.  No fun.
> ------------------------------------------
>
>                 Key: MAHOUT-690
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-690
>             Project: Mahout
>          Issue Type: Improvement
>    Affects Versions: 0.5
>            Reporter: Jake Mannix
>            Assignee: Jake Mannix
>             Fix For: 0.6
>
>         Attachments: MAHOUT-690.patch
>
>
> Long test runs long.
> Sample output:
> ...
> Running org.apache.mahout.math.hadoop.decomposer.TestDistributedLanczosSolver
> Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 282.859 sec
> ...
> This is very long, and should be improved, if possible.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira