Yes, you are right on that one ... I obviously did confuse the numbers. UDMA4 is indeed 66 MB/s. :o
But I wonder how you can measure the burst rate (which must be the IDE transfer rate peak out of your 716A's 8 MB buffer to your operating system's cache) so accurately without professional low-level measuring equipment since I have my doubts on synthetic benchmarking programs' accuracy.
Anyway, at 45 MB/s, the burst rate can be maintained for a maximum of 0,17 seconds (assuming that the complete 8192 kB of the 716A's drive cache are utilized which isn't the case IIRC) so this increase in speed (compared to 25 MB/s) will hardly be noticeable (if noticeable at all).
What's more important is the fact that the drive will reach its full continous speed (reading/writing) and acess time even in U-DMA2 ... at 33 MB/s the interface is still fast enough to maintain these speeds.
And I still keep asking myself if a VIA chipset (the Asus A7Nx series do feature nForce chipsets, not the A7Vx - my fault again) is the perfect partner for a high-speed optical device ... I have seen way too many unexplainable problems with VIA/nVidia/ALi chipsets (and drivers) in the past.