Hi I'm observing a huge difference in the performance speed of R on Windows and Unix, even though I know that my Unix machine is much more powerful than my Win machine. In particular, any character processing task is very time consuming on Unix. strptime(x,"%H:%M:%S") is about 10 times slower on Unix for vector x of the length of ~ 500. read.table() also is very slow. is there any way to speed up these ? Thanks a lot, Vlad.
morozov wrote:> Hi > I'm observing a huge difference in the performance speed of R on Windows and > Unix, even though I know that my Unix machine is much more powerful than my > Win machine. > In particular, any character processing task is very time consuming on Unix. > > strptime(x,"%H:%M:%S") is about 10 times slower on Unix for vector x of the > length of ~ 500. read.table() also is very slow. is there any way to speed up > these ? >As was pointed out, the strptime issue sounds like a C-library problem on your machine. The text processing might be the same. Is this a development version of a Un*x OS? What's the output of uname -a? Is it a source or binary build of R? If binary, try building and installing from source (provided you've got the disk space, you can do this, even if you're not root). Is the problem apparent on other boxen with similar OSs (ie is it hardware)? If you absolutely must continue to use that box and OS and R build, read.table() can be sped up using the colClasses argument (R doesn't have to guess what class each column should be), but it sounds like a problematic installation. Cheers Jason -- Indigo Industrial Controls Ltd. http://www.indigoindustrial.co.nz 64-21-343-545 jasont at indigoindustrial.co.nz
morozov wrote:>Hi >I'm observing a huge difference in the performance speed of R on Windows and >Unix, even though I know that my Unix machine is much more powerful than my >Win machine. >I've had this experience on a Unix machine that I thought was much more powerful than my Windows machine. On closer examination it seems the hardware was better at multitasking, but definitely not faster on single processes, so you might want to be careful about the part you think you know. I haven't had this experience running Windows and Linux on the same machine, but even then there are an enormous number of variable in the way you build R. Paul Gilbert>In particular, any character processing task is very time consuming on Unix. > >strptime(x,"%H:%M:%S") is about 10 times slower on Unix for vector x of the >length of ~ 500. read.table() also is very slow. is there any way to speed up >these ? > >Thanks a lot, >Vlad. > >______________________________________________ >R-help at stat.math.ethz.ch mailing list >https://www.stat.math.ethz.ch/mailman/listinfo/r-help > > >