Be aware that Xbench actually doesn't show the Mac Pro to be as good as it really is. If you actually watch your processor usage during the test they don't saturate. That means the test are actually over before the system has got itself going.
Additionally, the test is, rightfully or not, very heavily weighted to your drive performance. It is understandable given the largest bottleneck in any computer will be the hard drive. By upgrading your drive, more to the point having the right drive, you can hugely affect your scores. For some reason the 4K uncached write test seems to have a disporportionate affect on your drive score. My Maxtor MaXLines do extremely well in this test and for this alone they'll beat the Raptors in Xbench. A 2 x Raptor RAID 0 in actual fact beat my 3 x MaXLine RAID 0, sometimes very convincingly, in all other test yet will rate lower. So take these with a bit of a grain of salt.
As for the graphics processor, this barely makes a difference so if you have the ATi X1900 XT or Nvidia 7300 GT your scores will be essentially the same.
By the way, my 2.66GHz comes in at 218…
http://db.xbench.com/merge.xhtml?doc1=193878&doc2=1
… which show you that you can squeeze quite a bit more out of your 3GHz. By the way, my scores come from an every day work setup and not a benchmark setup like some others. Scores do seem to be better with Mac OS X 10.4.9 as well so be sure to be running this version. My score was done with 10.4.7.
Finally, Xbench is not always consistent so the best thing to do is to restart your MP, wait 30 seconds or so for it to settle down and then do like a best of 3 or so.