|Q: Efficient portable pseudo-codes email@example.com (1995-04-21)|
|From:||firstname.lastname@example.org (Tim Brabin Cooper)|
|Keywords:||interpreter, question, comment|
|Date:||Fri, 21 Apr 1995 16:05:50 GMT|
If you compile some high-level code down to pseudo-code, which
you then interpret, it'll be considerably slower than compiling
to straight native code. However, by choosing a good pseudo-code
instruction set, you can try to minimise this slow-down factor.
For example, you could create a single instruction for a common
idiom of instructions.
Has anyone done this, i.e. construct a pseudo-code instruction
set that can be interpreted very efficiently?
[Sure. Look at Smalltalk80, and at microcodable machines like the Burroughs
Return to the
Search the comp.compilers archives again.