[mlpack-svn] [MLPACK] #376: CosineTreeTest/CosineNodeCosineSplit test sometimes fails, only on i386

[MLPACK Trac]

#376: CosineTreeTest/CosineNodeCosineSplit test sometimes fails, only on i386
----------------------------------------------------+-----------------------
  Reporter:  rcurtin                                |        Owner:              
      Type:  defect                                 |       Status:  new         
  Priority:  minor                                  |    Milestone:  mlpack 1.1.0
 Component:  mlpack                                 |   Resolution:              
  Keywords:  cosine tree, cosinenode, cosine, test  |     Blocking:              
Blocked By:                                         |  
----------------------------------------------------+-----------------------

Comment (by siddharth.950):

 I noticed a problem with the test case, it uses LT(strict inequality)
 whereas the division of the node is not strict. However, looking at the
 values that is causing the test to fail, this is clearly not the issue.
 Also, the values don't seem to suggest a particular edge case that is
 being missed in the code. i386 is 32-bit, can it be an overflow(desperate
 speculation)? The description seems fine, unfortunately I cannot
 contribute more than this right now.

 This also reminds me of a problem I had noticed with mlpack tests. We
 randomly initialize the data matrices used to test the implementations,
 but in armadillo each time the 'random' values are the same. This
 practically makes the test use the same matrices again and again and we
 are less likely to encounter any flaws in the code/test. I think we should
 set the seed as you have done in the description and then initialize the
 matrices.

-- 
Ticket URL: <http://trac.research.cc.gatech.edu/fastlab/ticket/376#comment:1>
MLPACK <www.fast-lab.org>
MLPACK is an intuitive, fast, and scalable C++ machine learning library developed at Georgia Tech.