Skip to content

R with redhat 7 (PR#732)

2 messages · hyu@stats.uwo.ca, Peter Dalgaard

Original
hyu@stats.uwo.ca
# 
Full_Name: Hao Yu
Version: 1.1.0 and 1.1.1
OS: Redhat 7 Linux (intel)
Submission from: (NULL) (129.100.45.161)


Hi, I am not sure this is well know problem but I could not find anyone mention
it. 
The problem is related to build R on redhat 7. After got src.rpm and tried to
rebuild it, it failed on testing accuracy. Further it shows that under R,
.Machine gives some numbers that maybe are not correct. I also tried the most
recent one (November 6) and failed. The strange thing is that the R (1.1.0) that
comes with powertools of redhat 7 has exactly the same problem. All three R
gives the same results for .Machine. See the following outputs. 

Thanks
Hao Yu

PS. One system has Pentium III 600Mhz and 256RDRAM, Intel CC820 motherboard;
    Another has Pentium III 800Mhz and 256 SDRAM, ASUS BX motherboard.

***************************************************************************** 
R : Copyright 2000, The R Development Core Team
Version 1.1.1 Patched (November 6, 2000)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type    "?license" or "?licence" for distribution details.

R is a collaborative project with many contributors.
Type    "?contributors" for a list.

Type    "demo()" for some demos, "help()" for on-line help, or
        "help.start()" for a HTML browser interface to help.
Type    "q()" to quit R.
$double.eps
[1] 1.084202e-19

$double.neg.eps
[1] 5.421011e-20

$double.xmin
[1] 0.25

$double.xmax
[1] 0.5

$double.base
[1] 2

$double.digits
[1] 64

$double.rounding
[1] 5

$double.guard
[1] 0

$double.ulp.digits
[1] -63

$double.neg.ulp.digits
[1] -64

$double.exponent
[1] 1

$double.min.exp
[1] -2

$double.max.exp
[1] -2

$integer.max
[1] 2147483647

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Peter Dalgaard
# 
hyu@stats.uwo.ca writes:
We (R-core + Martyn Plummer who builds the RPM files on CRAN)
discussed this earlier today. It is a problem with the RedHat
"unofficical 2.96" gcc compiler which seems to be overly zealous on
some optimizations. The build is said to work if you build R
unoptimized.

We're strongly considering to say that R *requires* another gcc
compiler rather than trying to work around bugs in a compiler that
should never have been released.

The powertools version appears to have been built by disabling the
regression test (make check)! It might be worth filing a formal bug
report with RedHat on that matter. (It is not merely a matter of
cosmetics, those machine constants actually get used for determining
convergence criteria and the like. Having them set too low could send
algorithms into infinite loops.)

I think Martyn Plummer is about to produce a working RPM based on a
better compiler. However, if you want to compile add-on packages, it
might still be a good idea to install an alternative compiler for
yourself.