27 messages in com.perforce.jamming[jamming] Jam2.5 and precompiled headers
FromSent OnAttachments
Jacob Gorm Hansen22 Jan 2003 03:18 
Matt Armstrong22 Jan 2003 08:31 
Craig McPheeters22 Jan 2003 11:18 
Jacob Gorm Hansen23 Jan 2003 01:13 
Arnt Gulbrandsen23 Jan 2003 01:32 
Jacob Gorm Hansen23 Jan 2003 01:36 
Jacob Gorm Hansen23 Jan 2003 05:04 
Matt Armstrong23 Jan 2003 10:34 
Chris Antos23 Jan 2003 14:37 
Chris Antos23 Jan 2003 14:56 
Arnt Gulbrandsen24 Jan 2003 02:30 
Matt Armstrong24 Jan 2003 08:58 
Christopher Seiwald24 Jan 2003 16:50 
Matt Armstrong31 Jan 2003 15:04 
Chris Antos31 Jan 2003 15:34 
Elliot Murphy31 Jan 2003 18:58 
Matt Armstrong01 Feb 2003 08:55 
Elliot Murphy01 Feb 2003 09:37 
Chris Antos01 Feb 2003 12:02 
Craig McPheeters03 Feb 2003 10:35 
Jacob Gorm Hansen11 Feb 2003 02:46 
Jacob Gorm Hansen11 Feb 2003 04:56 
Craig McPheeters11 Feb 2003 10:29 
Chris Antos11 Feb 2003 11:13 
Matt Armstrong11 Feb 2003 22:45 
Chris Antos11 Feb 2003 23:27 
Matt Armstrong12 Feb 2003 09:46 
Subject:[jamming] Jam2.5 and precompiled headers
From:Christopher Seiwald (seiw@perforce.com)
Date:01/24/2003 04:50:05 PM
List:com.perforce.jamming

| > MSVC supports batching, and gets a big perf win from it. | | FWIW, gcc can do the same thing, but as near as I can tell it doesn't | help much. I tested just now and saw less than 1% speed improvement | (with about 40 C++ files). Disappointing.

That's largely why jam never did any such batching, because it didn't matter much at the time most of it was written. It's worth looking into the pch and batch optimizations, as header file mania has imposed a big penalty for a little "#include <windows.h>".

Christopher