How to get the Maximum or Minimum value of an array?

Comments

• pedward Posts: 1,642

  2011-12-24 02:26 edited 2011-12-24 02:26

  If you are talking about finding the largest or smallest number in an array, there is no builtin for that.
  
  You would have to do:
  

  pub max(arr, n)
  result := 0
  repeat x from 0 to n
    if arr[x] > result
      result := arr[x]
  
  pub min(arr, n)
  result := 0
  repeat x from 0 to n
    if arr[x] < result
      result := arr[x]
  
  
• kevinspace Posts: 56

  2011-12-24 02:34 edited 2011-12-24 02:34

  pedward,
  I am grateful for your assistance~
  
  
• MagIO2 Posts: 2,243

  2011-12-24 04:49 edited 2011-12-24 04:49

  @pedward:
  The idea is ok, but the implementation won't work!
  arr is a local variable that can be found on the stack. You can't pass arrays, only 32-bit values or pointers.
  arr[x] does not work, as it treats a variable like an array. arr[0] would in fact give you the content of arr, but arr[1] would give you the content of n. (the 2nd parameter on the stack) From there on you operate in memory areas which can contain anything.
  
  What you have to do is pass the address of the array to the function, which means that accessing the elements has to be changed.
  
  Additionally there are operators in SPIN which return the min and the max values, so you don't need to use the if-statement.
  

  pub max(arr_ptr, n)
    result :=0
    repeat x from 0 to n
      result #>= long[ arr_ptr ][ x ] 
  

  
  You'd call max like that:

  VAR
    long myArray[10]
  
  PUB main | maxVal
    maxVal := max( @myArray, 10 )
  

  
  PS: By the way ... there is another problem (in both implementations) if the numbers are allowed to be negative or the numbers are unsigned because SPIN deals with signed numbers. The negative number problem can be fixed easily by initializing result with the lowest possible negative number. Unsigned numbers need to be treated differently.
• Phil Pilgrim (PhiPi) Posts: 23,514

  2011-12-24 09:01 edited 2011-12-24 09:01

  I think both of you meant repeat x from 0 to n - 1, right? Also, you can't name the method max, since that's a reserved word, and don't forget to declare x. You can save one iteration by dong this:
  

  pub maxval(arr_ptr, n) | i
    result := long[arr_ptr]
    repeat i from 1 to n - 1
      result #>= long[arr_ptr][i]
  

  
  -Phil
• kwinn Posts: 8,697

  2011-12-24 10:00 edited 2011-12-24 10:00

  If you store the first element of the array in result rather than 0 it will handle negative numbers automatically. Also, you can find max_result and min_result with just one pass through the loop.
• MagIO2 Posts: 2,243

  2011-12-24 11:27 edited 2011-12-24 11:27

  Thanks for the pointers, that's why teamwork in most cases works better than single play ;o)
  
  Merry Xmas
• kuroneko Posts: 3,623

  2011-12-24 17:41 edited 2011-12-24 17:41

  pub maxval(arr_ptr, n) | i
    result := long[arr_ptr]
    repeat i from 1 to n - 1
      result #>= long[arr_ptr][i]
  

  long[base][idx] access is slow. This version - although slightly bigger - speeds things up a bit. Merry Christmas!

  pub maxval2(addr, n)
  
    addr := (addr - @addr) >> 2
    result := addr[addr++]
    repeat n - 1
      result #>= addr[addr++]