Hi, I get the wrong eigen values on an AIX machine with R-devel of 5/3/00.
Here are the results:
R> m <- matrix (c(6.8, 2.4, 2.4, 8.2), nrow=2)
R> m
[,1] [,2]
[1,] 6.8 2.4
[2,] 2.4 8.2
R> eigen(m)
$values
[1] 19.281403 6.337993
$vectors
[,1] [,2]
[1,] 0.1918866 0.9967987
[2,] 0.9967987 -0.1918866
And for comparison, here is what I get on a Sun (with r-stable):
R> m <- matrix (c(6.8, 2.4, 2.4, 8.2), nrow=2)
R> eigen (m)
$values
[1] 10 5
$vectors
[,1] [,2]
[1,] 0.6 0.8
[2,] 0.8 -0.6
I found this by running "make check" which also stumbles over svd.
Please let me know how I can help to investigate this.
-tom
--please do not edit the information below--
Version:
platform = powerpc-ibm-aix4.3.2.0
arch = powerpc
os = aix4.3.2.0
system = powerpc, aix4.3.2.0
status = Under development (unstable)
major = 1
minor = 1.0
year = 2000
month = May
day = 3
language = R
Search Path:
.GlobalEnv, Autoloads, package:base
--
mailto:tov@ece.cmu.edu (Tom Vogels) Tel: (412) 268-6638 FAX: -3204
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
tov@ece.cmu.edu writes:> Hi, I get the wrong eigen values on an AIX machine with R-devel of 5/3/00.Hmm. We had a similar report a while back and it turned out to be system specific. I believe the report was from Chong Gu.> Version: > platform = powerpc-ibm-aix4.3.2.0 > arch = powerpc > os = aix4.3.2.0 > system = powerpc, aix4.3.2.0 > status = Under development (unstable) > major = 1 > minor = 1.0 > year = 2000 > month = May > day = 3 > language = R-- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) 35327907 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel 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-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On Mon, 8 May 2000 tov@ece.cmu.edu wrote:> > Hi, I get the wrong eigen values on an AIX machine with R-devel of 5/3/00. >When we last saw this bug it was compiler-specific. gcc/g77 worked but cc/f77 didn't (and I think gcc/f77 wouldn't even configure). I don't think we understand the bug, though. -thomas -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel 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-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Hi, I ran the test again with the newly released version of R (1.1) and I
now get the correct eigen values. ("make check" succeeds.)
The scary part is that I think that I also used gcc+f77 last time and
the libs weren't changed. Maybe the scary part is that I think...
So the comments that this is compiler specific probably hold.
Thanks,
-tom
tov@ece.cmu.edu writes:
> Hi, I get the wrong eigen values on an AIX machine with R-devel of 5/3/00.
>
> Here are the results:
>
> R> m <- matrix (c(6.8, 2.4, 2.4, 8.2), nrow=2)
> R> m
> [,1] [,2]
> [1,] 6.8 2.4
> [2,] 2.4 8.2
> R> eigen(m)
> $values
> [1] 19.281403 6.337993
>
> $vectors
> [,1] [,2]
> [1,] 0.1918866 0.9967987
> [2,] 0.9967987 -0.1918866
>
>
> And for comparison, here is what I get on a Sun (with r-stable):
(same results now on AIX 4.3 with R 1.1.0
> R> m <- matrix (c(6.8, 2.4, 2.4, 8.2), nrow=2)
> R> eigen (m)
> $values
> [1] 10 5
>
> $vectors
> [,1] [,2]
> [1,] 0.6 0.8
> [2,] 0.8 -0.6
--
mailto:tov@ece.cmu.edu (Tom Vogels) Tel: (412) 268-6638 FAX: -3204
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._