Found this link to a Dell published PDF in Friday's blog.

It a very high level results comparison between various Dell servers. What is bothering me is that event though the hardware differs the performance (number of VMmark VMs) is comparable.

As a base line they use a 2950 dual dual core with 32gb of RAM. A 2900 with dual quad core and 48gb of RAM performs 51% better, this make a certain degree of sense to me.
What doesn't make sense is that the 6850 with quad dual cores and 64gb of RAM only performs 41% better, and similar, 57% for the 6950 with quad dual (AMD) and 64gb of RAM. I would have expected both these machines to be closer to twice the performance of the 2950.

Do these track with real world experience?

Side note: Unfortunately all the above configurations use 4gb memory sticks which still run at a very high premium. It would be interesting to see results from the above systems with the more affordable 2gb sticks.