compile time / run time - shift the borderline
in [C++]
|
Prev: pros and cons of returning const ref to string instead of string by value
Next: Updated performance tests using Boost Serialization lib
From: Frank Bergemann on 4 Dec 2009 11:15 Hi, reading about boost::proto i get back to the question of trading compile time efforts for run time efficiency. Unfortunately so far i didn't find a tutorial/example to hit the point here. In concrete i'd like to know, how run-time data go into the processing. I guess they have to move into what's names the "context" in tutorial of boost::proto(?) And the compiler can create a data-path structure for the context data - maybe pretty much like e.g. the butterfly-diagram for FFT-Calculation. Can anybody refer some good example for such? Another issue raising for studying boost::proto. Does it get us more closer to Compile-Time-Function-Evaluation (CTFE) for C++? So far templates is "programming with types". And we need to to use type-wrapper for numerics (e.g.) AFAIK it only supports calculations with built-in types and some basic operations for those. (e.g. there is no support for doubles and no support for thinks like sin(x)). I wonder what we would get, if we had the option for multi-level compilation respectively recursive compilation. Like (terminated) recursion is used with templates in one compilation "process". So why not allowing (recursive) sub-compilations? What's the problem in providing such for the compilers (like gcc)? I think it would be amazing options to have recursive code construction. It might set up another borderline shifting - this time between compilers and interpreters. cheers, Frank -- [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] |