Wait a minute!

As far as I have understood, "threaded" refers (in this context) to a 
style of writing (byte code) interpreters. A threaded interpreter does
not have a dispatch loop which reads the next byte code and then 
invokes the correct handler (typically by using the byte code as an index
into a table of handler addresses). Instead, the dispatch code (fetch
byte code, select handler) is integrated into each handler, eliminating
the return branch from the handler to the dispatch loop and allowing 
the overlap of the dispatch code with the handler code for much 
improved instruction scheduling.

Just my 2 �re.

/kff


_______________________________________________
Haskell mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/haskell

Reply via email to