From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00,FREEMAIL_FROM
	autolearn=ham autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,38fc011071df5a27
X-Google-Attributes: gid103376,public
X-Google-ArrivalTime: 2003-06-10 13:33:22 PST
Path: 
 archiver1.google.com!news1.google.com!newsfeed.stanford.edu!logbridge.uoregon.edu!snoopy.risq.qc.ca!chi1.webusenet.com!news.webusenet.com!news-feed01.roc.ny.frontiernet.net!nntp.frontiernet.net!uunet!dca.uu.net!ash.uu.net!spool.news.uu.net!not-for-mail
Date: Tue, 10 Jun 2003 16:33:10 -0400
From: Hyman Rosen <hyrosen@mail.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
 rv:1.4b) Gecko/20030603 Thunderbird/0.1a
X-Accept-Language: en-us, en
MIME-Version: 1.0
Newsgroups: comp.lang.ada
Subject: Re: Ideas for Ada 200X
References: <1055177817.528719@master.nyc.kbcfp.com>
 <Cm4Fa.80102$d51.148717@sccrnsc01>
In-Reply-To: <Cm4Fa.80102$d51.148717@sccrnsc01>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Organization: KBC Financial Products
Message-ID: <1055277191.38385@master.nyc.kbcfp.com>
Cache-Post-Path: master.nyc.kbcfp.com!unknown@nightcrawler.nyc.kbcfp.com
X-Cache: nntpcache 3.0.1 (see http://www.nntpcache.org/)
NNTP-Posting-Host: 204.253.250.10
X-Trace: 1055277191  1944 204.253.250.10
Xref: archiver1.google.com comp.lang.ada:38934
Date: 2003-06-10T16:33:10-04:00
List-Id: <comp.lang.ada>

tmoran@acm.org wrote:
 > how fast do they execute 10**7 times on 3x3 matrices of zeros?

Here's the fastest I've gotten so far.
Program output (times are in seconds):
+  8
+= 5
Compiler options:
     g++ -finline-limit=1000 -O3 -funroll-loops
Compiler version:
     g++.5.8 (GCC) 3.2.2
uname -a:
     SunOS 5.8 Generic_108528-15 sun4u sparc SUNW,Sun-Fire-880
Code:
#include <time.h>
#include <iostream>
#include <ostream>

template <typename L, typename R>
struct add_op
{
     const L &l;
     const R &r;
     add_op(const L &l, const R &r) : l(l), r(r) { }
     double operator()(int i, int j) const
     { return l(i,j) + r(i,j); }
};

template <typename L, typename R>
inline add_op<L,R> operator+(const L &l, const R &r)
{ return add_op<L,R>(l,r); }

template <int N>
struct Matrix
{
     double data[N][N];

     template <typename T>
     Matrix &operator=(const T &r)
     {
         for (int i = 0; i < N; ++i)
             for (int j = 0; j < N; ++j)
                 data[i][j] = r(i,j);
         return *this;
     }

     template <typename T>
     Matrix &operator+=(const T &r)
     {
         for (int i = 0; i < N; ++i)
             for (int j = 0; j < N; ++j)
                 data[i][j] += r(i,j);
         return *this;
     }

     double operator()(int i, int j) const
     { return data[i][j]; }
};

int main()
{
     Matrix<3> A, B;
     int i, j;
     for (i = 0; i < 3; ++i)
     {
         for (j = 0; j < 3; ++j)
         {
             A.data[i][j] = 0;
             B.data[i][j] = 0;
         }
     }

     long t1b = time(0);
     for (i = 0; i < 100000000; ++i) A = A + B;
     long t1e = time(0);
     long t2b = time(0);
     for (i = 0; i < 100000000; ++i) A += B;
     long t2e = time(0);
     std::cout << "+  " << t1e - t1b << '\n';
     std::cout << "+= " << t2e - t2b << '\n';
}