[Summary] Measuring the cost of synchronized I/O

1 view
Skip to first unread message

Alex Vinokur

unread,
Dec 30, 2004, 3:09:52 AM12/30/04
to C++ Perfometer SourceForge, C++ Perfometer Google Group


===================================
-------- Synchronized I/O ---------
Comparative performance measurement
===================================

Testsuite : Measuring the Cost of Synchronized I/O
Source : Technical Report on C++ Performance
Tool : The C/C++ Program Perfometer (Version 2.9.0-1.19)
* http://sourceforge.net/projects/cpp-perfometer/


Environment :
Windows 2000 Professional Ver 5.0 Build 2195 Service Pack 4
Intel(R) Celeron(R) CPU 1.70 GHz
CYGWIN_NT-5.0 1.5.12(0.116/4/2) i686 Cygwin
GNU g++ version 3.3.3 (cygwin special)

Compilation :
g++ *.cpp

DLLs : The names of DLL files on which the C++ Perfometer program depends
c:\cygwin\bin\cygwin1.dll
C:\WINNT\system32\ADVAPI32.DLL
C:\WINNT\system32\NTDLL.DLL
C:\WINNT\system32\KERNEL32.DLL
C:\WINNT\system32\RPCRT4.DLL



#=====================================================================
# ---------- Synchronized I/O ----------
# Measuring the Cost of Synchronized I/O
# ---- Summary ---
# ----------------
# ----------------------------------------
# The testsuite is from
# Technical Report on C++ Performance
# ISO/IEC PDTR 18015; Date: 2003-08-11; WG21 N1487=03-0070
# Appendix D: Timing Code
# SubAppendix D5. Measuring the Cost of Synchronized I/O
# ----------------------------------------
# * http://www.open-std.org/jtc1/sc22/WG21/docs/PDTR18015.pdf
# * http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2003/n1487.pdf
#---------------------------------------------------------------------
# Resource Name : CPU-time used
# Resource Cost Unit : clock
# Resource State Unit : clock
#=====================================================================
# Total repetitions : 100
# Performance metrics : clock / 100 repetitions
#=====================================================================


Measurement has been performed for the following optimization levels:
* No optimization

Total 100 repetitions
Performance = milliseconds per 100 repetitions


Here are summary results of the measurement.
--------------------------------------------

================================================
No optimization
------------------------------------------------
Full raw run log: http://groups-beta.google.com/group/log-files/msg/16d60f0b46fe5c9b
================================================

========================================================================
| | Size |
| Testsuite |-----------------------|
| | 100 : 1000 : 10000 |
|----------------------------------------------------------------------|
| integers to stdio | 33 : 290 : 4479 |
| hex integers to stdio | 40 : 347 : 3745 |
| integers to standard IO (sync = false) | 270 : 2747 : 28768 |
| hex integers to standard IO (sync = false) | 267 : 2704 : 27869 |
| integers to standard IO (sync = true) | 270 : 2750 : 28334 |
| hex integers to standard IO (sync = true) | 267 : 2710 : 27876 |
========================================================================


--
Alex Vinokur
email: alex DOT vinokur AT gmail DOT com
http://mathforum.org/library/view/10978.html
http://sourceforge.net/users/alexvn





--
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.298 / Virus Database: 265.6.6 - Release Date: 12/28/2004

Alex Vinokur

unread,
Dec 31, 2004, 1:20:36 PM12/31/04
to C++ Perfometer Google Group

===================================
-------- Synchronized I/O ---------
Comparative performance measurement
===================================

Testsuite : Measuring the Cost of Synchronized I/O
Source : Technical Report on C++ Performance
Tool : The C/C++ Program Perfometer (Version 2.9.0-1.19)
* http://sourceforge.net/projects/cpp-perfometer/


Environment :
Windows 2000 Professional Ver 5.0 Build 2195 Service Pack 4
Intel(R) Celeron(R) CPU 1.70 GHz
CYGWIN_NT-5.0 1.5.12(0.116/4/2) i686 Cygwin
GNU g++ version 3.3.3 (cygwin special)

Compilation :
g++ -O2 *.cpp
Optimization O2
------------------------------------------------
Full raw run log: http://groups-beta.google.com/group/log-files/msg/646d2d087efe31dc
================================================

========================================================================
| | Size |
| Testsuite |-----------------------|
| | 100 : 1000 : 10000 |
|----------------------------------------------------------------------|
| integers to stdio | 30 : 310 : 4112 |
| hex integers to stdio | 26 : 274 : 3578 |
| integers to standard IO (sync = false) | 217 : 2480 : 27062 |
| hex integers to standard IO (sync = false) | 217 : 2651 : 26605 |
| integers to standard IO (sync = true) | 210 : 2706 : 26611 |
| hex integers to standard IO (sync = true) | 217 : 2644 : 26625 |
========================================================================


--
Alex Vinokur
email: alex DOT vinokur AT gmail DOT com
http://mathforum.org/library/view/10978.html
http://sourceforge.net/users/alexvn







--
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.298 / Virus Database: 265.6.7 - Release Date: 12/30/2004

Reply all
Reply to author
Forward
0 new messages