Re: [R-SIG-Mac] test failures

2011-06-30 Thread Simon Urbanek
FWIW R built the official way doesn't have such problems, so I would not be surprised that it is an issue with your toolchain and/or the way you build R. Fink has been historically known for having issues with compilers (and library mismatches) so I would suggest looking at your end of the affa

Re: [R-SIG-Mac] test failures

2011-06-30 Thread Duncan Murdoch
On 11-06-30 9:48 PM, BABA Yoshihiko wrote: Dear all, I am a MacOSX/Fink (which is more or less like Debian's APT) maintainer and making R binary. One of the users tried test suite with make -k check and get the following results. Some differences are numerical, while others show sign differen

[R-SIG-Mac] test failures

2011-06-30 Thread BABA Yoshihiko
Dear all, I am a MacOSX/Fink (which is more or less like Debian's APT) maintainer and making R binary. One of the users tried test suite with make -k check and get the following results. Some differences are numerical, while others show sign differences in the result. Could someone possible ex

Re: [R-SIG-Mac] multicore package: collecting results

2011-06-30 Thread Simon Urbanek
On Jun 30, 2011, at 11:19 AM, Vincent Aubanel wrote: > > Le 30 juin 2011 à 15:36, Simon Urbanek a écrit : > >> >> On Jun 30, 2011, at 7:28 AM, Vincent Aubanel wrote: >> >>> Thanks for this, it's now dead fast, as one could conceivably expect. >>> Simon's solution is astonishingly fast, howeve

Re: [R-SIG-Mac] multicore package: collecting results

2011-06-30 Thread Vincent Aubanel
Le 30 juin 2011 à 15:36, Simon Urbanek a écrit : > > On Jun 30, 2011, at 7:28 AM, Vincent Aubanel wrote: > >> Thanks for this, it's now dead fast, as one could conceivably expect. >> Simon's solution is astonishingly fast, however I had to reconstruct the >> factors and their levels which were

Re: [R-SIG-Mac] multicore package: collecting results

2011-06-30 Thread Simon Urbanek
On Jun 30, 2011, at 7:28 AM, Vincent Aubanel wrote: > Thanks for this, it's now dead fast, as one could conceivably expect. > Simon's solution is astonishingly fast, however I had to reconstruct the > factors and their levels which were (expectedly) lost during the c() > operation. One way to

Re: [R-SIG-Mac] multicore package: collecting results

2011-06-30 Thread Vincent Aubanel
Thanks for this, it's now dead fast, as one could conceivably expect. Simon's solution is astonishingly fast, however I had to reconstruct the factors and their levels which were (expectedly) lost during the c() operation. Unfortunately this eats up some fair amount of cpu, but on a 14 columns, ~