We prefer to nvidia cards with CentOS-2. I have never had the problem of a monitor not being detected, though if you have tv out and no monitor plugged in then tv out will be used when the machine is powered on.

I have an RPM which I wrote which rebuilds the driver (if necessary) and loads it at boot time. Because it contains the nvidia package I don't know if I can redistribute it (have not checked) but here is my source part (Not tested with CentOS-3):

nvidia.spec =========== Summary: NVIDIA driver for XFree86 Name: nvidia Version: 5336 Release: 1 Vendor: Nvidia Distribution: Premonition Packager: John Newbigin jn@it.swin.edu.au License: Unknown Group: Applications/System BuildRoot: /var/tmp/%{name}-root Source0: NVIDIA-Linux-x86-1.0-%{version}-pkg1.run Source1: nvidia.rc

%description NVIDIA driver for XFree86

%prep if [ -d nvidia ] ; then rm -rf nvidia fi mkdir nvidia cd nvidia sh %{SOURCE0} -x

rm NVIDIA-Linux-x86-1.0-*/usr/src/nv/precompiled/*

%build

%install cd nvidia rm -rf ${RPM_BUILD_ROOT} mkdir -p ${RPM_BUILD_ROOT}/opt/nvidia cp -r NVIDIA-Linux-x86-1.0-*/* ${RPM_BUILD_ROOT}/opt/nvidia cp -r NVIDIA-Linux-x86-1.0-*/.manifest ${RPM_BUILD_ROOT}/opt/nvidia mkdir -p ${RPM_BUILD_ROOT}/etc/rc.d/init.d cp %{SOURCE1} ${RPM_BUILD_ROOT}/etc/rc.d/init.d/nvidia

%clean rm -rf $RPM_BUILD_ROOT

%files %defattr(-, root, root, 0755)

/etc/rc.d/init.d/nvidia /opt/nvidia/.manifest /opt/nvidia/*

%post if [ "$1" = "1" ] ; then chkconfig --add nvidia service nvidia start service nvidia configure else service nvidia stop service nvidia uninstall service nvidia start fi

%preun if [ "$1" = "0" ] ; then service nvidia stop service nvidia uninstall chkconfig --del nvidia fi

%changelog * Wed May 19 2004 John Newbigin jn@it.swin.edu.au - First Cut

nvidia.rc ========= #!/bin/bash # # chkconfig: 5 85 15 # description: Builds the nvidia kernel module for X11

# source function library . /etc/rc.d/init.d/functions

RETVAL=0

start() { NEED_BUILD=1 if [ -f /lib/modules/`uname -r`/kernel/drivers/video/nvidia.o ] ; then echo -n $"Loading nvidia driver" modprobe nvidia > /dev/null 2>/dev/null if grep -q nvidia /proc/modules ; then success $"Loading nvidia driver" echo NEED_BUILD=0 else failure $"Loading nvidia driver" echo fi fi

if [ $NEED_BUILD == 1 ] ; then echo -n $"Building nvidia driver" cd /opt/nvidia ./nvidia-installer --silent --no-precompiled-interface --no-network >/dev/null 2>/dev/null

if grep -q "is now complete" /var/log/nvidia-installer.log ; then success $"Building nvidia driver" echo echo -n $"Loading nvidia module" modprobe nvidia > /dev/null 2>/dev/null if grep -q nvidia /proc/modules ; then success $"Loading nvidia module" echo else failure $"Loading nvidia module" echo fi else failure $"Building nvidia driver" echo fi fi }

stop() { echo -n $"Unloading nvidia module: " rmmod nvidia > /dev/null 2>/dev/null # we don't care if it does not work...? success RETVAL=$?

echo } configure() { echo -n "Configuring XFree86-4 for nvidia"

if [ ! -f /etc/X11/XF86Config-4.prenvidia ] ; then cp /etc/X11/XF86Config-4 /etc/X11/XF86Config-4.prenvidia fi #replace the line: # Driver "nv" # (or Driver "vesa") # #with # # Driver "nvidia" # #In the Module section, make sure you have: # # Load "glx" # #You should also remove the following lines: # # Load "dri" # Load "GLcore"

/usr/bin/perl -pi -e "s|Driver[[:space:]]*"nv"|Driver "nvidia"|" /etc/X11/XF86Config-4 if ! grep -q "^[[:space:]]*Load[[:space:]]*"glx"" /etc/X11/XF86Config-4 ; then echo "We need to add Load "glx"" fi if grep -q "^[[:space:]]*Load[[:space:]]*"dri"" /etc/X11/XF86Config-4 ; then /usr/bin/perl -pi -e "s|Load[[:space:]]*"dri"|#Load "dri"|" /etc/X11/XF86Config-4 fi if grep -q "^[[:space:]]*Load[[:space:]]*"GLcore"" /etc/X11/XF86Config-4 ; then /usr/bin/perl -pi -e "s|Load[[:space:]]*"GLcore"|#Load "GLcore"|" /etc/X11/XF86Config-4 fi success echo }

uninstall() { cd /opt/nvidia ./nvidia-installer --silent --uninstall }

case "$1" in start) start ;; stop) stop ;; configure) configure ;; uninstall) uninstall ;; *) echo $"Usage: $0 {start|stop|configure|uninstall}" exit 1 esac

exit $RETVAL

C.M. Connelly wrote:

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

An option to installing the nVidia drivers with nVidia's installer is to get the dell-nvidia and dkms packages from Dell's support website (search for dell-nvidia and dkms). The best thing about this approach is that the driver will be recompiled when the machine is rebooted with a new kernel.

As a (perhaps irrelevant) data point, I'm running a PowerEdge 400SC with an nVidia Corporation GeForce3 Ti 200 and a DVI connection to a Dell 2000FP monitor at the full 1600x1200 resolution. Works great with the nVidia drivers; only the analog connection works for X with the XFree driver.

I have a GeForce4 card with the same monitor at home in my Debian system; works great with the Debian-packaged nVidia drivers.

Claire

*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* Claire Connelly cmc@math.hmc.edu Systems Administrator (909) 621-8754 Department of Mathematics Harvey Mudd College *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) Comment: Processed by Mailcrypt 3.5.8 http://mailcrypt.sourceforge.net/

iD8DBQFBUb57B0pE8d7vd8wRAobXAJ9fLysDccEY3nn4f+xPkA3ZeEFGMgCggz80 vhQ1XyztGmoIG7s7w3Vo9kY= =/1P5 -----END PGP SIGNATURE----- _______________________________________________ CentOS mailing list CentOS@caosity.org http://www.caosity.org/mailman/listinfo/centos