dvanbrunt@well-wired.com
2004-May-21 15:24 UTC
[Rd] Memory Leak in OS X versions? (PR#6903)
Full_Name: David L. Van Brunt Version: 1.8-1.9 beta OS: OS X 10.3 Submission from: (NULL) (68.74.58.109) As posted on R-Help (after which another user replicated the problem): --------------- This is the conclusion from a prior thread ([R] " cannot allocate vector of length 1072693248") which ended with no other answer but that there must be a problem in the OS X version of R, or in the compile of the source on OS X. I?ve posted code and data here: http://www.well-wired.com/reflibrary/uploads/1084503247.zip If you setwd() into the directory that is made, then ?source()? the ?.R? file, it should run fine on Windows but crash on any machine with OS X (Panther) giving: ?Error in as.vector(data) : cannot allocate vector of length 1073741824? after a few iterations of the loop. I've repeated this on a Powerbook G4 with 500 MB of RAM, an iMac with 128M of RAM, and a Dual 2GHz G5 with 1.5 Gig of RAM. Have used the Raqua, the Frameworks installation, and a fresh compile of the source using Fink in the X11 implementation. No matter the machine or the version (1.8x through 1.9, OS X or Unix X11), I get the same result. Thanks to Andy Liaw for helping me tune the code this far. I'm stumped. Would love to hear others' experience with this, and if they can reproduce the problem elsewhere. -- David L. Van Brunt, Ph.D. Outlier Consulting & Development mailto: <ocd@well-wired.com>
I can confirm the behaviour (btw, 1.9.0 is not beta) in 1.9.0 and in R-patched I'll try to investigate. stefano On May 21, 2004, at 3:24 PM, dvanbrunt@well-wired.com wrote:> Full_Name: David L. Van Brunt > Version: 1.8-1.9 beta > OS: OS X 10.3 > Submission from: (NULL) (68.74.58.109) > > > As posted on R-Help (after which another user replicated the problem): > --------------- > This is the conclusion from a prior thread ([R] " cannot allocate > vector of > length 1072693248") which ended with no other answer but that there > must be > a problem in the OS X version of R, or in the compile of the source on > OS X. > > I?ve posted code and data here: > http://www.well-wired.com/reflibrary/uploads/1084503247.zip > > If you setwd() into the directory that is made, then ?source()? the > ?.R? > file, it should run fine on Windows but crash on any machine with OS X > (Panther) giving: ?Error in as.vector(data) : cannot allocate vector of > length 1073741824? after a few iterations of the loop. > > I've repeated this on a Powerbook G4 with 500 MB of RAM, an iMac with > 128M > of RAM, and a Dual 2GHz G5 with 1.5 Gig of RAM. Have used the Raqua, > the > Frameworks installation, and a fresh compile of the source using Fink > in the > X11 implementation. No matter the machine or the version (1.8x through > 1.9, > OS X or Unix X11), I get the same result. > > Thanks to Andy Liaw for helping me tune the code this far. > > I'm stumped. Would love to hear others' experience with this, and if > they > can reproduce the problem elsewhere. > > > -- > David L. Van Brunt, Ph.D. > Outlier Consulting & Development > mailto: <ocd@well-wired.com> > > ______________________________________________ > R-devel@stat.math.ethz.ch mailing list > https://www.stat.math.ethz.ch/mailman/listinfo/r-devel >
Reasonably Related Threads
- Memory Leak in OS X version of R?
- Can't seem to finish a randomForest.... Just goes and goe s!
- " cannot allocate vector of length 1072693248"
- Can't seem to finish a randomForest.... Just goes and goes!
- Repost: Examples of "classwt", "strata", and "sampsize" i n randomForest?