Some of your goals are unrealistic/unachievable. An example is the 10 millisecond field to field transition time. Most Monitors in desktop use refresh at the 60hz rate, which leaves each frame on the screen for 16-17ms. Some studies suggest that the barrier for human perceptibility is about 50ms. So, to even test this you would need specialized hardware and software that most of your users are not going to have deployed due to cost.
The remainder of the items are easily covered by a standard automated functional testing tool with built in timing capability. Even though automated functional tools have built in timers (well, most of the tools anyway), few people seem to take advantage of this capability for timing and thus collection of performance information earlier in the QA cycle, instead of leaving questions of performance to the "performance team."
Your last two items can be accomplished using a headless database query tool which returns the first byte of the returned data set. Some automated functional testing tools have the capability of connecting to a database directly to run such queries, but not all. Setting a timer is key here as well. You can also pull a profiling report to cross check the performance of a particular set of queries across a variety of parameter sets.