The following graph shows the relative performances for a number of different scenarios (with Dynamic VHD being the baseline).
Fixed VHD always performs better than a Dynamic VHD in most scenarios by roughly 10% to 15% with the exception of 4k writes, where Fixed VHD performs significantly better. More
I believe your table is based on the VHD performance study published by Microsoft in Nov 209 and than updated in early 2010.
It is important to remember that performance data in that study was based on FULLY expanded DVHD, in which case one would reasonably expect its performance to be roughly the same as fixed VHD. So if anything these numbers demonstrate an extremely POOR implementation of the DVHD in the R2 releases,.
The more practical case of partially expanded DVHD performance is not covered at all. In my own tests partially expanded DVHD perform anywhere between 2.5 times (40%) worse than fixed (for fully random write loads) to about the same for pure read loads (obviously!).
IMHO even Fixed VHD is not an adequate solution for high performance virtualization with Hyper-V. You may check our product Virsto One (www.virsto.com) which makes Hyper-V a true enterprise class virtualization solution and enables it to compete with VMware on a more equal footing.
alex at virsto dot com