www.digitalmars.com         C & C++   DMDScript  

c++ - why are c++ template instantiations inherently slow?

reply evansl <cppljevans suddenlink.net> writes:
In this post:

  http://article.gmane.org/gmane.comp.lib.boost.devel/189925

Eric indicates that Walter Bright believes, in c++:

  that instantiating a template is inherently expensive, and certain
  features of the C++ language (ADL, partial specialization, etc.)
  force that to be the case.

Walter, if Eric is remembering correctly, could you provide a little
more explanation on why this is so or provide a link to some document
supporting this conclusion?

TIA.

-regards,
Larry
Oct 09 2012
next sibling parent reply Nick Sabalausky <SeeWebsiteToContactMe semitwist.com> writes:
On Tue, 09 Oct 2012 15:59:26 -0500
evansl <cppljevans suddenlink.net> wrote:

 In this post:
 
   http://article.gmane.org/gmane.comp.lib.boost.devel/189925
 
 Eric indicates that Walter Bright believes, in c++:
 
   that instantiating a template is inherently expensive, and certain
   features of the C++ language (ADL, partial specialization, etc.)
   force that to be the case.
 
 Walter, if Eric is remembering correctly, could you provide a little
 more explanation on why this is so or provide a link to some document
 supporting this conclusion?
 
 TIA.
 
 -regards,
 Larry

"C++ Compilation Speed" by Walter Bright: http://www.drdobbs.com/cpp/c-compilation-speed/228701711 Note the #7.
Nov 01 2012
parent evansl <cppljevans suddenlink.net> writes:
On 11/01/12 04:29, Nick Sabalausky wrote:
 On Tue, 09 Oct 2012 15:59:26 -0500
 evansl <cppljevans suddenlink.net> wrote:
 
 In this post:

   http://article.gmane.org/gmane.comp.lib.boost.devel/189925

 Eric indicates that Walter Bright believes, in c++:

   that instantiating a template is inherently expensive, and certain
   features of the C++ language (ADL, partial specialization, etc.)
   force that to be the case.

 Walter, if Eric is remembering correctly, could you provide a little
 more explanation on why this is so or provide a link to some document
 supporting this conclusion?

 TIA.

 -regards,
 Larry

"C++ Compilation Speed" by Walter Bright: http://www.drdobbs.com/cpp/c-compilation-speed/228701711 Note the #7.

I read it and now I'm heavily leaning toward converting to D now. However, I saw no mention of ADL and partial specialization. Is there some other reference explaining why these lead to slow compile times? -regards, Larry
Nov 03 2012
prev sibling parent Nick Sabalausky <SeeWebsiteToContactMe semitwist.com> writes:
On Sat, 03 Nov 2012 12:36:00 -0500
evansl <cppljevans suddenlink.net> wrote:

 On 11/01/12 04:29, Nick Sabalausky wrote:
 
 "C++ Compilation Speed" by Walter Bright:
 http://www.drdobbs.com/cpp/c-compilation-speed/228701711
 
 Note the #7.
 

I read it and now I'm heavily leaning toward converting to D now. However, I saw no mention of ADL and partial specialization. Is there some other reference explaining why these lead to slow compile times?

Not that I know of, but I would assume it's because (IIUC) those are heavily template-oriented things which, in turn, are slow in C++ because of the reasons he described. You could try asking over on the "digitalmars.D" newsgroup, as that tends to get a lot more traffic (same NNTP server as here, ie, news.digitalmars.com).
Nov 03 2012