Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.compilers > #775

basic question about cps

From n.oje.bar@gmail.com
Newsgroups comp.compilers
Subject basic question about cps
Date 2012-11-12 11:05 -0800
Organization Compilers Central
Message-ID <12-11-004@comp.compilers> (permalink)

Show all headers | View raw


Hi, I have a basic question about writing a compiler using continuation-passing style
as an intermediate language.

It seems to me that if one performs the cps transformation one is left
with many 'computed' calls (that is call to variables holding
procedure values, namely, the continuation). It seems that compiling
such 'computed' calls would be much slower than compiling 'direct'
calls to known procedures.

Question:

1. Is this assumption valid?
2. If it is, then it seems that cps is not very useful without some
non-trivial control flow analysis...??

Please excuse me if this question is very basic.

Thanks!
Nicolas
[Yes, it assumes a reasonbly good optimizer. -John]

Back to comp.compilers | Previous | NextNext in thread | Find similar


Thread

basic question about cps n.oje.bar@gmail.com - 2012-11-12 11:05 -0800
  Re: basic question about cps torbenm@diku.dk (Torben Ægidius Mogensen) - 2012-11-14 11:25 +0100
  Re: basic question about cps Stefan Monnier <monnier@iro.umontreal.ca> - 2012-11-14 11:13 -0500
    Re: basic question about cps n.oje.bar@gmail.com - 2012-11-21 09:27 -0800

csiph-web