[R] Corrections for Solaris stuff in manual

Dr. David Kirkby david.kirkby at onetel.net
Mon May 31 19:16:51 CEST 2010


There are a number of errors in the R manual about Solaris.

http://cran.r-project.org/doc/manuals/R-admin.html#Solaris

1) Firstly, Sun are now owned by Oracle, who bought them for $7 billion.

2) "(Recent Sun machines are Opterons (‘amd64’) rather than ‘x86’, but 32-bit 
‘x86’ executables are the default.) "

That's incorrect. All recent Sun workstations used *Intel* CPUs - Opterons have 
not been used for some time. The directory structure is however the same as with 
the operterons (using /usr/lib/amd64).

I'm using a Sun Ultra 27 (quad core 3.33 GHz Xeon) which I bought about 6 months 
ago. That is a current model. Last time I looked, the Ultra 24 was the lower-end 
x64 box, and again used some sort of Intel CPUs.

3) "Modern Solaris systems allow a large selection of Open Source software to be 
installed from http://www.opencsw.org (formerly http://www.blastwave.org) via 
pkg-get, "

Again incorrect. Both Blastwave and OpenCSW are in existence. There is a lot of 
bad feelings between the two camps, but they do both still exist.


I was going to ask if you knew how to build R as a 64-bit application on 
OpenSolaris x64, though I see you indicate

############
Tests with gcc34 on ‘x86’ and ‘amd64’ have been less successful: ‘x86’ builds 
have failed on tests using complex arithmetic35, whereas on ‘amd64’ the builds 
have failed to complete in several different ways, most recently with relocation 
errors for libRblas.so.
############

I get that when trying to build R within Sage:

http://trac.sagemath.org/sage_trac/ticket/9040

ld.so.1: R: fatal: relocation error: R_AMD64_PC32: file 
/export/home/drkirkby/sage-4.4.2/spkg/build/r-2.10.1.p1/src/lib/libR.so: symbol 
_init: value 0x228000984acd does not fit

There's some info about this problem at

http://developers.sun.com/solaris/articles/about_amd64_abi.html

where it is suggests -Kpic is used to build Position Independent Code (PIC) 
code. I've not tried using -fPIC with gcc, which might solve the problem, as 
that builds PIC code.

Dave



More information about the R-help mailing list