Simple code performance improvement in C

Asked

Viewed 103 times

3

I had a proof that the question was :

Cosiderando that the periods of history are four:

(1) period prior to 4000 BC;

(2) from 3999 BC to 476 AD;

(3) from 477 AD to 1789 AD; and

(4) after 1790 AD.

Note that BC corresponds to before Christ, and AD corresponds to after Christ, and the birthmark of Christ is the year 0 (zero).

Write function that has as input parameters, the year and reference as to the Christian era, and return the number corresponding to the respective period of history: 1,2, 3 or 4. Then read 3 pairs of year and reference to the Christian era and identify the most recent (largest). the output format is : 3 pairs of year and reference to the Christian era, each in a separate line.

I did the code on the test and then submitted it to a website where I was evaluating the execution time, and other students did the same. My code made in C was not the fastest, others did in C and had a better execution time, so I would like to know how to improve the execution time of this code, where I am missing, if there is any redundancy, to be able to make better codes, the code is as follows:: 

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

int main(){

char buffer1[8],buffer2[8],buffer3[8],buffer[8];

int Ano,i,space,Anos[3],cont=0;

char letra,letras[4];

char st_int[5];

//--------------------------------------------------------------------------
gets(buffer1);
gets(buffer2);          //RECEBE OS DADOS
gets(buffer3);
//--------------------------------------------------------------------------
 //------------------------------------------------------------------------   
inicio:
if(cont==0)strcpy(buffer,buffer1);
else if(cont==1)strcpy(buffer,buffer2);
else strcpy(buffer,buffer3);
for(i=0;i<8;i++){
space=buffer[i];                            //TRATA OS DADOS DA ENTRADA
if(space == 32){
    letra=buffer[i+1];
    st_int[i]='\0'; Ano=atoi(st_int);//printf("Ano=!%d!\n",Ano);
    break;
    }
st_int[i]=buffer[i];
}

if(cont<=2){
    letras[cont]=letra; Anos[cont]=Ano;
cont++;
if(cont<=2)goto inicio;
}
letras[cont]='\0';
//--------------------------------------------------------------------------
//--------------------------------------------------------------------------
for(i=0;i<=2;i++){
    //printf("entrou\n");
    if(letras[i] == 'a'){
        if(Anos[i]>4000){
            Anos[i]=1;continue;
            }                     //VERIFICA EM QUAL ERA OS ANOS SE ENCAIXAM
        if(Anos[i]<=3999){
            Anos[i]=2;continue;            
        }
    }
    if(letras[i]=='d'){
        if((Anos[i]>=0) && (Anos[i]<=476)){
            Anos[i]=2;continue;
        }
        if((Anos[i]>=477) && (Anos[i]<=1789)){
            Anos[i]=3;continue;
        }
        if(Anos[i]>=1790){
            Anos[i]=4;continue;
        }
    }
    }
//------------------------------------------------------------------------
//------------------------------------------------------------------------
if(Anos[0]>Anos[1])
    cont=Anos[0];                     //VERIFICA QUAL O MAIOR VALOR DA ERA
else
    cont=Anos[1];
if(Anos[1]>Anos[2])
    cont=Anos[1];
else
    cont=Anos[2];
//------------------------------------------------------------------------
printf("\n%d",cont);  //MOSTRA O RESULTADO FINAL

system("PAUSE");
return 0;
    }

I appreciate any hint/criticism/advice to improve the performance of this code and the next ones I will do.

c performance
  • 4

    Well, for starters, all calls from strcpy spend precious time unnecessarily. Try to declare buffer as char *buffer and make buffer = buffer3 (for example) instead of copying. You’ll already see some improvement. Another thing: the line with letra=buffer[i+1]; has a potential memory invasion, when the i is equal to 7 (that is, if by any chance there is a space in the last position).

  • 4

    When you say the performance was worse, how many percent are we talking about? I didn’t see anything screaming inefficient in your code at most I think you could avoid the 3 strcpy using references/pointers instead (have you studied pointers?). The rest is micro-optimization, I don’t know if it’s worth heating up with this. By the way, that system("PAUSE") in the end does not interfere with the execution time analysis, certain?

  • I hadn’t thought about this way of using pointers to replace strcpy that you talked about, although I already know pointers and have used it in some cases without even realizing that I could do it. It was really helpful, thank you.

1 answer

2

Follows a very efficient solution to the problem:

#include <stdio.h>
#include <stdlib.h>
#include <string.h>


int parse_data( const char * str )
{
    int ano;

    ano = atoi(str);

    if( ano == 0 )
        return 2;

    if( strchr( str, 'a' ) )
    {
        if( ano > 4000 )
            return 1;
        else    
            return 2;
    }

    if( strchr( str, 'b' ) )
    {
        if( ano <= 476 )
            return 2;

        if( (ano >= 477) && (ano <= 1789) )
            return 3;

        if( ano > 1789 )
            return 4;
    }

    return -1;
}


int main( int argc, char * argv[] )
{
    int ret;
    char data[ 16 ];

    printf("Data: ");
    fgets( data, sizeof(data), stdin );

    ret = parse_data( data );

    printf("\nCodigo: %d\n", ret );

    return 0;
}

/* fim-de-arquivo */

I hope it helps!

  • I thought it was good how you did it in such a succinct way, but how the function strchr() does not decrease the performance since it is making another function call?

Browser other questions tagged c performance

You are not signed in. Login or sign up in order to post.