Re: c++/2183: gcc-3_0-branch compiles 2x slower than 2.95

Re: c++/2183: gcc-3_0-branch compiles 2x slower than 2.95

[bangerth]

Synopsis: gcc-3_0-branch compiles 2x slower than 2.95

State-Changed-From-To: open->closed
State-Changed-By: bangerth
State-Changed-When: Fri Nov  1 17:40:57 2002
State-Changed-Why:
    I think a compile-time regression in this range must be
    expected due to the switch to libstdc++ v3 alone. Add some
    new optimizations, and you're there, so I think one should
    classify this as "more features make things slower".
    
    Needless to say that I do not find the general trend too
    good either...

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=2183

c++/2183: gcc-3_0-branch compiles 2x slower than 2.95

[gisli]

>Number:         2183
>Category:       c++
>Synopsis:       gcc-3_0-branch compiles 2x slower than 2.95
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Mar 05 10:06:02 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     gisli@adams.com
>Release:        CVS gcc-3_0-branch on March 3rd
>Organization:
>Environment:
System: Linux 2.2.16-22smp
Architecture: i686
host: i686-pc-linux-gnu
build: i686-pc-linux-gnu
target: i686-pc-linux-gnu
configured with: configure --prefix=/opt/gcc3.0 --enable-shared

>Description:
This is my first experience with GCC 3.0.  I am compiling
a ca. 100,000 line C++ program which takes modest advantage
of templates (STL mostly) and exceptions.  By virtue of
being ported to a variety of platforms, the code compiled
with 3.0 pretty much out of the box.

However, compile time is noticably higher, going from 
exactly 15 minutes with 2.95 to exactly 30 minutes with 3.0.

Pertinent compile flags are:  -ggdb -ansi -pedantic -W

I have not compared the compile time for optimization,
because this does not affect my daily work.

I am available to do any kind of testing/experiments that
you would like me to.  If you are interested, I should be
able to provide pre-processed source, e.g. for the following
file:

AsExpression.o

2.95, -ggdb -ansi -pedantic -W              25.02s
3.0-branch, -ggdb -ansi -pedantic -W        44.24s

2.95, -O2                                  207.28s 
3.0-branch, -O2                            392.45s

Unfortunately, due to a lack of a binary compatible 
support library, I have not yet been able to test the
performance of the resulting executable.

>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: