[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-02-08 Thread jkbradley
Github user jkbradley commented on the issue:

https://github.com/apache/spark/pull/16495
  
Thanks @mhmoudr
As far as the stress test, I'd recommend posting instructions as a Github 
gist and linking it to wherever you post results on JIRA or a PR.  We wouldn't 
want to add a model (a binary file) to git unless absolutely necessary.

When we address the complexity issue, we can post stress test results and 
the link to the gist.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-02-08 Thread mhmoudr
Github user mhmoudr commented on the issue:

https://github.com/apache/spark/pull/16495
  
* To traing a (one-off) model a keep it in resources, then load it to run 
the test.
* investigate the added comlexity issue.
* To move the test away from unit tests
* Provide instructions on how to run the stress test


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-02-08 Thread mhmoudr
Github user mhmoudr commented on the issue:

https://github.com/apache/spark/pull/16495
  
Sure .. I am, but just a bit busy nowadays, in general I am planing to 
investigate the additional complexity issue, and as for the test I agree with 
the fact that the test should not be added to unit test as far as there is a 
specific agreed manner for adding stress tests, also to add, in term of the 
test itself, I would say it will be a wast of time to train the model  every 
time we need to test this area, so I was thinking of train a model once and 
save it to resources, then once I need to test the evaluateEachIteration it 
will be much faster to lead the pre saved model from resources and run the test 
directly, thoughts?

for the time being I will close the pull request and once I have the change 
ready I will create a new one.
Mahmoud


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-02-06 Thread jkbradley
Github user jkbradley commented on the issue:

https://github.com/apache/spark/pull/16495
  
@mhmoudr Will you be able to update this?


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-02-04 Thread vlad17
Github user vlad17 commented on the issue:

https://github.com/apache/spark/pull/16495
  
Yes, sorry for my wording. A unit test is indeed an inappropriate place for 
stress tests. An offline test would be sufficient to verify that an O(N) 
implementation is an improvement over the O(N^2) one. Ideally the stress test 
would be neatly described in the PR message so that anyone could replicate, 
perhaps even linking to a gist with the script to run it.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-01-09 Thread jkbradley
Github user jkbradley commented on the issue:

https://github.com/apache/spark/pull/16495
  
The key point @vlad17 made was that an operation which should be O(N) is 
taking O(N^2) in the current implementation.  Let's fix that, regardless of 
whether or not we add a stress test.

Also, we can't add such a long test for this.  Unit tests should take 
seconds at most, and ideally sub-second times.  Offline stress testing is 
great, though.

Thanks!


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-01-06 Thread AmplabJenkins
Github user AmplabJenkins commented on the issue:

https://github.com/apache/spark/pull/16495
  
Merged build finished. Test PASSed.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-01-06 Thread AmplabJenkins
Github user AmplabJenkins commented on the issue:

https://github.com/apache/spark/pull/16495
  
Test PASSed.
Refer to this link for build results (access rights to CI server needed): 
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/71003/
Test PASSed.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-01-06 Thread SparkQA
Github user SparkQA commented on the issue:

https://github.com/apache/spark/pull/16495
  
**[Test build #71003 has 
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/71003/testReport)**
 for PR 16495 at commit 
[`d3e2dad`](https://github.com/apache/spark/commit/d3e2dadaa767bd3fec10ba329625ceaa5ccabcbb).
 * This patch passes all tests.
 * This patch merges cleanly.
 * This patch adds no public classes.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #16495: SPARK-16920: Add a stress test for evaluateEachIteration...

2017-01-06 Thread SparkQA
Github user SparkQA commented on the issue:

https://github.com/apache/spark/pull/16495
  
**[Test build #71003 has 
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/71003/testReport)**
 for PR 16495 at commit 
[`d3e2dad`](https://github.com/apache/spark/commit/d3e2dadaa767bd3fec10ba329625ceaa5ccabcbb).


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org