Well, having read this thread, I'd have to say that overall, I'll put my eggs in your basket. I really don't blame the people who try to make a buck by selling all kinds of cures for diseases that don't exist. But it should be kept in mind that there are smart people in Redmond who have a vested interest in making Windows run as well and as fast as possible out of the box. I've tried multiple "registry tools" over the years, some seemingly doing some good, some making my machines unbootable. I still use only one - Auslogics Registry Defrag. Not sure that is does anything at all, but gives that "feelgood" feeling. An analogy I use sometimes about the registry in order to disabuse people of the notion that its size or degree of "fragmentation" has any bearing on access speed is to point out that if you go to a restaurant, you are led to the table that you will use. If the other tables in the room are filled to the brim or not, your speed in getting to your table is unaffected because you don't start at the first one and move through them all until you get to yours. And if the room gets bloated by having an addition added on, that has no effect, either. That is the way binary databases work. But, for those who really want to spend time on this, there does exist a registry benchmarking tool. The speed changes are extremely small, generally. And, it's a work-in-progress. http://www.bitsum.com/regbench.php (It should be noted that the system hive is loaded to RAM by the bootloader, right after the kernel, the HAL, and vgaoem.fom. The hive then has random access and there is no difference in speed between different locations' access time.)