>But, if I had to boot the standard kernel, is there a way I could >install the NVidia drivers from there? How do I tell them to compile >against the lowlat CCRMA kernel? How do I tell them to install to >the appropriate place? It doesn't seem like this would be terribly >difficult. I just don't know what parameters/environment variables >to change. > >Thanks, >Greg > > > from the nvidia documentation : ..... Since the Linux kernel does not support a binary driver interface, we provide for rebuilding these files on the target machine (or distribution) and then linking with the binary version of the NV kernel driver. $ make This will produce one of the following files depending on the system include directory used to do the build: NVdriver: which is the installable kernel driver module for single processor Linux systems If you want to build NVdriver for a system other than the compiling system, then you'll need to run the make as: $ make SYSINCLUDE=/src/kern/my-smp-kernel/include to generate an NVdriver that will work on the kernel whose include files are in /src/kern/my-smp-kernel/include. This kernel must have been completely configured (make menuconfig dep). ..... i hope this helps -- rossen