[R] help with debugging hardware dependent problems

[Clayton Springer]

Dear r-help,

I am having an intermittant error using R on a linux beowulf cluster.

I am using my homebrew script to test the performance of various feature subsets
with "lda".

-------------

The general procedure is as follows:

I load the data from an .Rdata file (which has generated by R-1.3.0)

run my script "Iterate()" ...

At a different iteration each time I get the error:

Error in svd(X, nu = 0) : NA/NaN/Inf in foreign function call (arg 1)

It is as if the data is not properly loaded and R just puts "NA"
instead.

The error is inconsistent and not repeatable.  Even when conditions are 
repeated it doesn't happen at the same iteration everytime.

The error does not occur at all on my stand alone linux box.

--------------

Can anyone suggest debugging tools to track the problem more 
precisely?

In both cases, I compiled the R-1.4.0 myself. The Beowulf cluster is RedHat linux and
my stand alone box is Mandrake.

The Beowulf cluster is using a RAID, are there any known issues with those?


Thanks for your help,


--Clayton



----------


Here is an actual transcript. In which the model is run twice, but dies in
a different place 
It shows that the "Iterate()" is dieing on different
iterations. traceback() shows that it is lda having the problem. 

>  Iterate(1,200)
1 thesefactors  V2 
2 thesefactors  V3 
3 thesefactors  V2 V3 
4 thesefactors  V4 
5 thesefactors  V2 V4 
6 thesefactors  V3 V4 
Error in svd(X, nu = 0) : NA/NaN/Inf in foreign function call (arg 1)
> Iterate(1,200)
1 thesefactors  V2 
2 thesefactors  V3 
3 thesefactors  V2 V3 
4 thesefactors  V4 
5 thesefactors  V2 V4 
6 thesefactors  V3 V4 
7 thesefactors  V2 V3 V4 
8 thesefactors  V5 
9 thesefactors  V2 V5 
10 thesefactors  V3 V5 
11 thesefactors  V2 V3 V5 
12 thesefactors  V4 V5 
13 thesefactors  V2 V4 V5 
14 thesefactors  V3 V4 V5 
15 thesefactors  V2 V3 V4 V5 
Error in svd(X, nu = 0) : NA/NaN/Inf in foreign function call (arg 1)
> traceback()
5: svd(X, nu = 0)
4: lda.default(x, grouping, ...)
3: lda.formula(fmla, data = data)
2: lda(fmla, data = data)
1: Iterate(1, 200)


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._