public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug libstdc++/32618]  New: std::vector calls uneccessary constructors instead of inplace construction of first object
@ 2007-07-04  0:46 oliver at kfs dot org
  2007-07-04  0:52 ` [Bug libstdc++/32618] " pinskia at gcc dot gnu dot org
                   ` (11 more replies)
  0 siblings, 12 replies; 13+ messages in thread
From: oliver at kfs dot org @ 2007-07-04  0:46 UTC (permalink / raw)
  To: gcc-bugs

When using resize() and push_back(), an baseline instace of T is created and
then the new members are assigned via the copy constructor.

If T is complex this begins to add a non-negligible overhead, especially for
push_back.

Is there a reason /not/ to replace the [first] create & copy with in-place
construction?

Sample code:

  #include <iostream>

  struct Foo {
   Foo() { cout << "ctor " << this << endl ; }
   Foo(const Foo& rhs) { cout << "copyctor " << this << " from " << &rhs <<
endl ; }
   ~Foo() { cout << "dtor " << this << endl ; }
  } ;

  typedef vector<Foo> FOO ;

  int main(int argc, char* argv[]) {
   FOO bar;
   cout << "pushing" << endl ;
   bar.push_back(Foo()) ;
   cout << "end of program" << endl ;
  }

Compilation:
  [osmith@fedex src]$ g++ -v -save-temps -O2 -g2 -Wall -o test -lstdc++
test.cpp
  Using built-in specs.
  Target: i386-redhat-linux
  Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada --enable-java-awt=gtk
--disable-dssi --enable-plugin
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre --with-cpu=generic
--host=i386-redhat-linux
  Thread model: posix
  gcc version 4.1.1 20070105 (Red Hat 4.1.1-51)
   /usr/libexec/gcc/i386-redhat-linux/4.1.1/cc1plus -E -quiet -v -D_GNU_SOURCE
test.cpp -mtune=generic -Wall -fworking-directory -O2 -fpch-preprocess -o
test.ii
  ignoring nonexistent directory
"/usr/lib/gcc/i386-redhat-linux/4.1.1/../../../../i386-redhat-linux/include"
  #include "..." search starts here:
  #include <...> search starts here:
   /usr/lib/gcc/i386-redhat-linux/4.1.1/../../../../include/c++/4.1.1
  
/usr/lib/gcc/i386-redhat-linux/4.1.1/../../../../include/c++/4.1.1/i386-redhat-linux
   /usr/lib/gcc/i386-redhat-linux/4.1.1/../../../../include/c++/4.1.1/backward
   /usr/local/include
   /usr/lib/gcc/i386-redhat-linux/4.1.1/include
   /usr/include
  End of search list.
   /usr/libexec/gcc/i386-redhat-linux/4.1.1/cc1plus -fpreprocessed test.ii
-quiet -dumpbase test.cpp -mtune=generic -auxbase test -g2 -O2 -Wall -version
-o test.s
  GNU C++ version 4.1.1 20070105 (Red Hat 4.1.1-51) (i386-redhat-linux)
          compiled by GNU C version 4.1.1 20070105 (Red Hat 4.1.1-51).
  GGC heuristics: --param ggc-min-expand=98 --param ggc-min-heapsize=129076
  Compiler executable checksum: 4720743fdfefd64206c8550433f6e508
   as -V -Qy -o test.o test.s
  GNU assembler version 2.17.50.0.6-2.fc6 (i386-redhat-linux) using BFD version
2.17.50.0.6-2.fc6 20061020
   /usr/libexec/gcc/i386-redhat-linux/4.1.1/collect2 --eh-frame-hdr -m elf_i386
--hash-style=gnu -dynamic-linker /lib/ld-linux.so.2 -o test
/usr/lib/gcc/i386-redhat-linux/4.1.1/../../../crt1.o
/usr/lib/gcc/i386-redhat-linux/4.1.1/../../../crti.o
/usr/lib/gcc/i386-redhat-linux/4.1.1/crtbegin.o
-L/usr/lib/gcc/i386-redhat-linux/4.1.1 -L/usr/lib/gcc/i386-redhat-linux/4.1.1
-L/usr/lib/gcc/i386-redhat-linux/4.1.1/../../.. -lstdc++ test.o -lstdc++ -lm
-lgcc_s -lgcc -lc -lgcc_s -lgcc /usr/lib/gcc/i386-redhat-linux/4.1.1/crtend.o
/usr/lib/gcc/i386-redhat-linux/4.1.1/../../../crtn.o

Output:
  pushing
  ctor 0xbfa3d327
  copyctor 0x8c87008 from 0xbfa3d327
  dtor 0xbfa3d327
  end of program
  dtor 0x8c87008

If resize were being called, it the copyctor would be used to copy /0xbfa3d327/
into each new uninitialized instance:
  resizing
  ctor 0xbfa3d327
  copyctor 0x8c87008 from 0xbfa3d327
  copyctor 0x8c8700c from 0xbfa3d327
  copyctor 0x8c87010 from 0xbfa3d327
  dtor 0xbfa3d327
  end of program
  dtor 0x8c87008
  dtor 0x8c8700c
  dtor 0x8c87010


The optimization I am suggesting would produce the following [approximate]
output:
  pushing
  ctor 0x8c87008
  end of program
  dtor 0x8c87008

If resize() was being called instead of push_back, it might look like this:
  resizing
  ctor 0x8c87008   /first entry in the pool/
  copyctor 0x8c8700c from 0x8c87008 /copy ctor of second entry from first/
  copyctor 0x8c87010 from 0x8c87008 /copy ctor of third entry from first/
...
  end of program
  dtor 0x8c87008
  dtor 0x8c8700c
  dtor 0x8c87010


-- 
           Summary: std::vector calls uneccessary constructors instead of
                    inplace construction of first object
           Product: gcc
           Version: 4.1.1
            Status: UNCONFIRMED
          Severity: minor
          Priority: P3
         Component: libstdc++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: oliver at kfs dot org
  GCC host triplet: i386-redhat-linux
GCC target triplet: i386-redhat-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32618


^ permalink raw reply	[flat|nested] 13+ messages in thread
end of thread, other threads:[~2010-06-18 18:22 UTC | newest]

Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-07-04  0:46 [Bug libstdc++/32618] New: std::vector calls uneccessary constructors instead of inplace construction of first object oliver at kfs dot org
2007-07-04  0:52 ` [Bug libstdc++/32618] " pinskia at gcc dot gnu dot org
2007-07-04  1:03 ` pcarlini at suse dot de
2007-07-04  4:27 ` oliver at kfs dot org
2007-07-04  4:34 ` pinskia at gcc dot gnu dot org
2007-07-04  5:45 ` oliver at kfs dot org
2007-11-28  0:57 ` pcarlini at suse dot de
2008-01-02 20:37 ` pcarlini at suse dot de
2009-12-02  0:01 ` paolo dot carlini at oracle dot com
2010-01-08 15:42 ` paolo dot carlini at oracle dot com
2010-06-16 11:04 ` paolo dot carlini at oracle dot com
2010-06-18 18:08 ` paolo at gcc dot gnu dot org
2010-06-18 18:22 ` paolo dot carlini at oracle dot com

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).