R help - Aug 2004 - hclust-segmentation fault

I am getting the "Segmentation fault" when using hclust in R-1.9.1
running
under SuSe 9.0 64-bit kernel on a dual opteron system with 8G of RAM. 
I was wandering if anybody could offer any insight?
Thanks,
mario.

>>>>> "MarioM" == Medvedovic, Mario (medvedm)
<medvedm at UCMAIL.UC.EDU>
>>>>>     on Mon, 9 Aug 2004 10:47:16 -0400  writes:
    MarioM> I am getting the "Segmentation fault" when using
    MarioM> hclust in R-1.9.1 running under SuSe 9.0 64-bit
    MarioM> kernel on a dual opteron system with 8G of RAM.  I
    MarioM> was wandering if anybody could offer any insight?

{ where will you get when "wandering" around?  :-) }

I have seen other problematic behavior on a 64-bit kernel Linux
on the Opteron though not with SuSE and only with R-devel (R-2.0.0-unstable)

Can you give a reproducible example please
(as the posting guide asks you to)?

Martin Maechler

On Mon, 9 Aug 2004, Medvedovic, Mario (medvedm) wrote:
> I am getting the "Segmentation fault" when using hclust in
R-1.9.1 running
> under SuSe 9.0 64-bit kernel on a dual opteron system with 8G of RAM. 
> I was wandering if anybody could offer any insight?
Please try to use the debugger to supply more information, or give us some 
code we can reproduce on a similar system to see if we can reproduce the 
segfault.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

Well, the use of debugger will take some time, but here is a simple code
that invariably causes the fault. 
Mario.

indata<-matrix(rnorm(1000,0,1),ncol=10)
ed<-dist(indata)
hc.e<-hclust(ed,"average")

>-----Original Message-----
>From: Prof Brian Ripley [mailto:ripley at stats.ox.ac.uk]
>Sent: Monday, August 09, 2004 11:14 AM
>To: Medvedovic, Mario (medvedm)
>Cc: 'r-help at stat.math.ethz.ch'
>Subject: Re: [R] hclust-segmentation fault
>
>
>On Mon, 9 Aug 2004, Medvedovic, Mario (medvedm) wrote:
>
>> I am getting the "Segmentation fault" when using hclust in 
>R-1.9.1 running
>> under SuSe 9.0 64-bit kernel on a dual opteron system with 
>8G of RAM. 
>> I was wandering if anybody could offer any insight?
>
>Please try to use the debugger to supply more information, or 
>give us some 
>code we can reproduce on a similar system to see if we can 
>reproduce the 
>segfault.
>
>-- 
>Brian D. Ripley,                  ripley at stats.ox.ac.uk
>Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
>University of Oxford,             Tel:  +44 1865 272861 (self)
>1 South Parks Road,                     +44 1865 272866 (PA)
>Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>

Thanks a lot for the assistance. It was an installation mistake - apparently
we did not have BLAS libraries installed. The compiler did not seem to have
been a problem in this case, althought it is an older version (gcc 3.3.1).
Mario.

>-----Original Message-----
>From: Prof Brian Ripley [mailto:ripley at stats.ox.ac.uk]
>Sent: Monday, August 09, 2004 11:51 AM
>To: Medvedovic, Mario (medvedm)
>Cc: 'r-help at stat.math.ethz.ch'
>Subject: RE: [R] hclust-segmentation fault
>
>
>On Mon, 9 Aug 2004, Medvedovic, Mario (medvedm) wrote:
>
>> Well, the use of debugger will take some time, but here is a 
>simple code
>> that invariably causes the fault. 
>> Mario.
>> 
>> indata<-matrix(rnorm(1000,0,1),ncol=10)
>> ed<-dist(indata)
>> hc.e<-hclust(ed,"average")
>
>Works fine on R 1.9.1 on our dual Opteron 248 under FC2.
>
>We know of some pertinent compiler bugs on x86_64, so is this 
>gcc 3.3.3 
>or later?
>
>> >-----Original Message-----
>> >From: Prof Brian Ripley [mailto:ripley at stats.ox.ac.uk]
>> >Sent: Monday, August 09, 2004 11:14 AM
>> >To: Medvedovic, Mario (medvedm)
>> >Cc: 'r-help at stat.math.ethz.ch'
>> >Subject: Re: [R] hclust-segmentation fault
>> >
>> >
>> >On Mon, 9 Aug 2004, Medvedovic, Mario (medvedm) wrote:
>> >
>> >> I am getting the "Segmentation fault" when using
hclust in
>> >R-1.9.1 running
>> >> under SuSe 9.0 64-bit kernel on a dual opteron system with 
>> >8G of RAM. 
>> >> I was wandering if anybody could offer any insight?
>> >
>> >Please try to use the debugger to supply more information, or 
>> >give us some 
>> >code we can reproduce on a similar system to see if we can 
>> >reproduce the 
>> >segfault.
>> >
>> >-- 
>> >Brian D. Ripley,                  ripley at stats.ox.ac.uk
>> >Professor of Applied Statistics, 
http://www.stats.ox.ac.uk/~ripley/
>> >University of Oxford,             Tel:  +44 1865 272861 (self)
>> >1 South Parks Road,                     +44 1865 272866 (PA)
>> >Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>> >
>> 
>> ______________________________________________
>> R-help at stat.math.ethz.ch mailing list
>> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide! 
>http://www.R-project.org/posting-guide.html
>> 
>> 
>
>-- 
>Brian 
>D. Ripley,                  ripley at stats.ox.ac.uk
>Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
>University of Oxford,             Tel:  +44 1865 272861 (self)
>1 South Parks Road,                     +44 1865 272866 (PA)
>Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>