[R] configuring graphics device in Linux
john polo
jpolo at mail.usf.edu
Sat Nov 15 14:08:50 CET 2008
megha patnaik wrote:
> Hi John,
> an R problem, but an X11 bug in Ubuntu. This is evident from
plot(rnorm(10))
> failing to perform. If you are seeing this, then it is because X11
does not
> automatically detect Intel graphic cards. The solution is to edit your
> xorg.conf file, changing it to something like (this is from my xorg.conf):
>
> Section "Device"
> Identifier "Intel Corporation Mobile 945GM/GMS, 943/940GML Express
> Integrated Graphics Controller (rev 03)"
> Driver "intel"
> EndSection
Megha, thanks for your help.
i tried "plot(rnorm(10))" and still was not able to see a plot. i
checked /etc/X11/xorg.conf and it looks like this, without me editing it:
Section "Device"
Identifier "Intel Corporation Mobile 945GM/GMS, 943/940GML
Express Integrated Graphics Controller"
Driver "intel"
BusID "PCI:0:2:0"
EndSection
Section "Monitor"
Identifier "Generic Monitor"
Option "DPMS"
EndSection
Section "Screen"
Identifier "Default Screen"
Device "Intel Corporation Mobile 945GM/GMS, 943/940GML
Express Integrated Graphics Controller"
Monitor "Generic Monitor"
DefaultDepth 24
SubSection "Display"
Modes "1280x800"
EndSubSection
EndSection
> The steps I carried out to detect where the problem was were as follows:
> 1. Checked whether basic plots in R worked
> 2. Created a new UID ( which did not have my saved settings in R) and
tried
> basic plots.
>
> If 2. told me that margins were too large, then this was not a problem due
> to my past commands in R.
i have tried your suggested plot and the one from the tutorial. neither
worked. i'm using R 2.6.2 straight from the install through Unbuntu's
package manager, so i don't think there are any scripts or whathaveyou
on start.
> xorg.conf is a dangerous file to meddle with. The other symptoms I was
> seeing of poor graphics was an XL font size on my login screen and broken
> images in google earth. Are you seeing anything of the sort?
yes, actually i do get the large fonts at start up like you do. i tried
trolling around the Ubuntu forums for stuff about video settings, but
only briefly.
> I am relatively new to the mailing list. Do tell me if this is better
taken
> offline.
i'm kind of new too. maybe someone else will tell me(us) to take this
somewhere else.
Peter Dalgaard wrote:
> Good advice in general, but in this case I think John is right in
assuming that the issue is mainly with his graphics hardware setup, and
he did tell us about that. The R version would have >been nice though;
we have fixed various issues in this area over time.
thank you Peter. i appreciate you assistance. it's version 2.6.2
> According to your setup, the system should believe that you have an
>~8 inch screen (800/96) and >the default X11 window is 7x7 inches so
>there _should_ be plenty of room, but some systems try >to outsmart
>the user and use the actual physical dimensions, which could be
>smaller (higher >DPI). You might want to check with "xdpyinfo" (from
>the shell).
from xdpyinfo:
screen #0:
dimensions: 1280x800 pixels (289x21 millimeters)
resolution: 112x968 dots per inch
> In any case, you can try reducing width, height, and (especially)
pointsize in the call to x11() (which >you'd issue _before_ the plot()).
Or >set them with X11.options().
i tried playing with plot(),halving (and again) the w,h, and pointsize,
to no avail. and help.search("X11.options") reveals nothing.
More information about the R-help
mailing list