Well, for sure, you can only run 1024 QAM on "bluebird days," but that's the point of adaptive modulation. Although having this sort capability that can only be used under the right conditions would seem like a huge waste in the consumer electronics space, infrastructure is a completely different space.
It would be interesting to know how often the system can actually operate in 1024 QAM mode in the real world. In a pure AWGN channel, 1024 QAM requires "only" four times as much power as 256 QAM at the same BER, but the situation quickly gets much worse in the presence of interference.
The 25% capacity gain over 256 QAM is probably well worth the higher electric bill, but what is the effective capacity gain on a typical long-haul system? Can an operator expect to operate in 1024 QAM mode 25% of the time, 75% of the time, or what?
Blog Doing Math in FPGAs Tom Burke 24 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...