writing byte per byte is relative slow, check the overhead of loop() itself and the call to write.
did also some testing with higher baud rates and used a bit longer strings, see - Fast communication to PC with low Arduino footprint - #12 by robtillaart - Networking, Protocols, and Devices - Arduino Forum -
The longer strings reduces the overhead substantially.
Also at the receiving end I used an terminal program called putty.exe. These terminals (there are others too) are quite optimized for communication, where Processing uses probably a layer of libraries (assumption)