.text
    .align 4
    .globl procesar
    .type   procesar, @function
procesar:
    pushl   %ebp
    movl    %esp, %ebp
    subl    $16, %esp
    pushl   %ebx
    pushl   %esi
    pushl   %edi

# Aqui has de introducir el codigo
  xorl %edi, %edi # edi <- i = 0
  movl 16(%ebp), %esi # esi = N
  imul %esi, %esi # esi = N*N

  movl 8(%ebp), %ecx # ecx = *mata
  movl 12(%ebp), %edx # edx = *matb

  #for 0 -> n^2
bucle:
    cmpl %esi, %edi
  jge fifor

  movdqu (%ecx, %edi), %xmm0
  paddb %xmm0, %xmm0 # mata[i*n+j]*2
  paddb %xmm0, %xmm0 # mata[i*n+j]*4
  paddb %xmm0, %xmm0 # mata[i*n+j]*8
  paddb %xmm0, %xmm0 # mata[i*n+j]*16
  movdqu %xmm0, (%edx, %edi) # matb[i*n+j] <- mata[i*n+j]*16

  addl $16, %edi #add 16
  jmp bucle
fifor:

# El final de la rutina ya esta programado

    emms    # Instruccion necesaria si os equivocais y usais MMX
    popl    %edi
    popl    %esi
    popl    %ebx
    movl %ebp,%esp
    popl %ebp
    ret