It's all about reverse-engineering stuff; my usual workflow looks like: (1) Complile some Swift program & look at the disasembler listing of it (I use IDA and — for Objective-C — Hopper Disassembler); (2) Do some changes in source code in the program, compile it again, and see diffs in disassembler listing; that way I can see what is changed and how.

About the @asmname thing: you can easily dump every default Swift module using a REPL command:

repl> :print_module Swift

There you can find a lot of interesting language lexems and more, so all you have to do is to try using them in your code until you don't figure it out ;)