Using the 'nosmp' cmdline option probably won't help at all... and it's VERY old information, since the kernel has been autodetecting multiple cores (or just one!) on it's own for a while now.
I'd REALLY recommend using a newer version of ANY Linux distribution. RHEL 5.1 was released in 2007, there've been major changes and improvements, and using something that old is pretty much setting yourself up for headaches later, it's likely to have many vulnerabilities which have been fixed/patched since then... (6 years!!)
RHEL 6.4 was released in February of this year (2013), it would be the best thing if you were going to stick with RHEL. Of course, there's the price.
Fedora is a viable (affordable!) option, same RedHat, just without the prepaid support plan. Besides, you come here first, just like the rest of you guys (I see you lurking!).
Off the cuff list: Slackware, Mint (very popular lately), Ubuntu, Debian (my favorite), CentOS ("Enterprise Quality" they say), OpenSuSE.. I've lost track. They use CentOS last place I worked for all their servers. I use Debian 7 (Jessie!) for all mine at home.
It boils down to this. Obtain a newer release of Linux, whatever flavor you desire, it should work very well. The RHEL 5.1 may not recognize some device in the newer machine which is making udev lock up (although I do recall a problem with udev a few years ago... but it was fixed) (see! new release!)
Now, IF you were to try to boot that system with 'nosmp'... and IF it successfully booted and was 'usable'... you'd be running a multi-core machine with tons of power... off of a single core. That'd be like driving a V8 engined car... but only using 2 cylinders (a V2?)... and wondering why it hasn't got any pickup.
But due to the udev issue you've mentioned, I don't think that 6 year old Linux likes the new devices it found when udev tried to populate /dev. Confused it's little brain.
I'm Ramblin' again...