#include

#include

#define TILE$\_$SIZE $16$

#define UNROLL$\_$FACTOR $8$

void matrix$\_$multiplication$($int $*$A$,$ int $*$B$,$ int $*$C$,$ int N$)\{$

$//$ Initialize result matrix C to $0$

for $($int i $=0$; i $<$ N $*$ N; $++$i$)\{$

C$[$i$]=0$;

$\}$

$//$ Iterate over the tiles

for $($int i $=0$; i $<$ N; i $+=$ TILE$\_$SIZE$)\{$

for $($int j $=0$; j $<$ N; j $+=$ TILE$\_$SIZE$)\{$

for $($int k $=0$; k $<$ N; k $+=$ TILE$\_$SIZE$)\{$

$//$ Compute the block sub$-$matrix

for $($int ii $=$ i; ii $<$ i $+$ TILE$\_$SIZE && ii $<$ N; $++$ii$)\{$

for $($int jj $=$ j; jj $<$ j $+$ TILE$\_$SIZE && jj $<$ N; $++$jj$)\{$

int sum $=0$;

int kk;

for $($kk $=$ k; kk $<$ k $+$ TILE$\_$SIZE && kk $+($UNROLL$\_$FACTOR $-1)<$ N; kk $+=$ UNROLL$\_$FACTOR$)\{$

$//$ Prefetching next elements

$\_\_$builtin$\_$prefetch$($&A$[$ii $*$ N $+$ kk $+$ UNROLL$\_$FACTOR$],0,3)$;

$\_\_$builtin$\_$prefetch$($&B$[($kk $+$ UNROLL$\_$FACTOR$)*$ N $+$ jj$],0,3)$;

$//$ Loop unrolling

sum $+=$ A$[$ii $*$ N $+$ kk$]*$ B$[$kk $*$ N $+$ jj$]$;

sum $+=$ A$[$ii $*$ N $+$ kk $+1]*$ B$[($kk $+1)*$ N $+$ jj$]$;

sum $+=$ A$[$ii $*$ N $+$ kk $+2]*$ B$[($kk $+2)*$ N $+$ jj$]$;

sum $+=$ A$[$ii $*$ N $+$ kk $+3]*$ B$[($kk $+3)*$ N $+$ jj$]$;

sum $+=$ A$[$ii $*$ N $+$ kk $+4]*$ B$[($kk $+4)*$ N $+$ jj$]$;

sum $+=$ A$[$ii $*$ N $+$ kk $+5]*$ B$[($kk $+5)*$ N $+$ jj$]$;

sum $+=$ A$[$ii $*$ N $+$ kk $+6]*$ B$[($kk $+6)*$ N $+$ jj$]$;

sum $+=$ A$[$ii $*$ N $+$ kk $+7]*$ B$[($kk $+7)*$ N $+$ jj$]$;

$\}$

$//$ Handle the remaining elements

for $($; kk $<$ k $+$ TILE$\_$SIZE && kk $<$ N; $++$kk$)\{$

sum $+=$ A$[$ii $*$ N $+$ kk$]*$ B$[$kk $*$ N $+$ jj$]$;

$\}$

C$[$ii $*$ N $+$ jj$]+=$ sum;

$\}$

$\}$

$\}$

$\}$

$\}$

$\}$

can sopmebody modify this withoput changing its functionality