Sorry, I am against of all this.
Increase of Tx power. Three reasons why I am against of this:
1. Influence to health.
2. Really it will not increase the speed and range, because you have also increase the Tx power of your clients. Otherwise you will get rather speed or/and range degradation. Just client can "hear" your router, but your router does not "hear" your client and repeats again what was already said.
3. It is some kind of law violations (crime). What could be problematic not only for you, but for me too if I provide such possibility.
Moreover, FCC regulations do not allow this, so manufacturers lock down this possibility. Such stuffs are hard-coded in pre-built parts of manufactures drivers, but not in open codes.
Usually increase of Wi-Fi range is performed by use of directional antennas in concrete direction. Or by adding special re-translators (additional access points). You can try to use different country codes with your router, e.g. set US or TW if you are not from these countries but e.g. from Europe. Just to test will it be useful for your clients.
OC: Also I am against of this. AL-514 is already one of the most powerful Cortex-A15. Netgear developers were unable to use all potential of this CPU, because their firmware is optimized for generic ARM-V7a. I.e. they do not use full set if CPU instructions. But my build does that, because is optimized for concrete Cortex-A15. Plus, increase of CPU clocking does not help in such basic stuffs as throughput of usual traffic: e.g. hardware acceleration of NAT is a work for chipset and corresponding drivers but not for CPU. CPU is used for such stuffs as e.g. OpenVPN with its encryption. But for this much more correct approach is hardware acceleration of OpenSSL. E.g. my version of OpenSSL works is in 6 times faster in some encryption operation than stock version. So you would not get such acceleration just overclocking CPU but still using stock firmware.
Plus possible damage of unit...
Voxel.