Date: Wed, 4 Mar 1998 16:42:59 -0500 (EST) Message-Id: <199803042142.QAA02942@p2.acadia.net> To: Bum-Seok Hyun Subject: Re: [Q]Computing speed in C++ Cc: djgpp AT delorie DOT com References: <34FCB769 DOT 42BEF1A8 AT gong DOT snu DOT ac DOT kr> <199803040910 DOT EAA08076 AT p2 DOT acadia DOT net> <34FD2F86 DOT E3DCCB5D AT gong DOT snu DOT ac DOT kr> in-reply-to: <34FD2F86.E3DCCB5D@gong.snu.ac.kr> From: swarnerx3 AT acadia DOT net (Scott Warner) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit Precedence: bulk Interesting! AFAIK, djgpp is a good optimizing compiler, but there may be something funny happening with your in-loop declarations. (Although my guess is that gxx is gcc with different libraries.) The program may have to continuously reallocate the heap (if that's the precise term). I suggest trying the identical code in C and C++, so you're comparing apples and apples. > > > The test code is quite simple, consisted of 3 nested "for" loops. > > > > Let me first thank you for your kind concern. > > Main body of my test code is > > for(int i=1 ; i<10001 ; i++) > for(int j=1 ; j<10001 ; j++) > for(int k=1 ; k < 11 ; k++) sum += i; > > Above code is ,of course , for C++. > > In C code, I just moved variable definitions > into the first line of the code. > Like, > > int i,j,k; > float sum; > > That's all. > > I don't think that makes musch difference between C++ and C. > > And compile option is , > > C code : gcc -O2 -o testc.exe testc.c > C++ code : gxx -O2 -o testcpp.exe testcpp.cpp > > And I'm using DJGPP v.2.10 > > Did I make some mistakes or misunderstand something important? > > Anyway, thank you and wish you goooooooooooooooooooood Luck !!!! > > Bum-Seok Hyun > mailto:dominic AT gong DOT snu DOT ac DOT kr