PureBasic Forums - English

Posted: **Wed Jul 10, 2024 2:46 pm**

SMaag wrote: Wed Jul 10, 2024 1:50 pm I tried to calculate the BayerMatrix

This one uses recursion

Global Dim M(15, 15)

Procedure InitBayer(x, y, s, v, st)
  If s = 1
    M(y, x) = v
  Else
    s >> 1
    InitBayer(x,   y,   s, v,       st<<2)
    InitBayer(x+s, y+s, s, v+st,    st<<2)
    InitBayer(x+s, y,   s, v+st<<1, st<<2)
    InitBayer(x,   y+s, s, v+st*3,  st<<2)
  EndIf
EndProcedure

Debug "DataSection"
Debug ""
s = 2
While s <= 16
  InitBayer(0, 0, s, 0, 1)
  Debug "  ; "+Str(s)+"x"+Str(s)+" Bayer Matrix"
  Debug "  Bayer"+Str(s)+"x"+Str(s)+":"
  For y = 0 To s-1
    ds.s = "  Data.a "
    For x = 0 To s-2
      ds + RSet(Str(M(y, x)),3)+", "
    Next
    ds + RSet(Str(M(y, s-1)),3)
    Debug ds
  Next
  Debug "  Bayer"+Str(s)+"x"+Str(s)+"_end:"
  Debug ""
  s << 1
Wend
Debug "EndDataSection"

Output:

Code: Select all

DataSection

  ; 2x2 Bayer Matrix
  Bayer2x2:
  Data.a   0,   2
  Data.a   3,   1
  Bayer2x2_end:

  ; 4x4 Bayer Matrix
  Bayer4x4:
  Data.a   0,   8,   2,  10
  Data.a  12,   4,  14,   6
  Data.a   3,  11,   1,   9
  Data.a  15,   7,  13,   5
  Bayer4x4_end:

  ; 8x8 Bayer Matrix
  Bayer8x8:
  Data.a   0,  32,   8,  40,   2,  34,  10,  42
  Data.a  48,  16,  56,  24,  50,  18,  58,  26
  Data.a  12,  44,   4,  36,  14,  46,   6,  38
  Data.a  60,  28,  52,  20,  62,  30,  54,  22
  Data.a   3,  35,  11,  43,   1,  33,   9,  41
  Data.a  51,  19,  59,  27,  49,  17,  57,  25
  Data.a  15,  47,   7,  39,  13,  45,   5,  37
  Data.a  63,  31,  55,  23,  61,  29,  53,  21
  Bayer8x8_end:

  ; 16x16 Bayer Matrix
  Bayer16x16:
  Data.a   0, 128,  32, 160,   8, 136,  40, 168,   2, 130,  34, 162,  10, 138,  42, 170
  Data.a 192,  64, 224,  96, 200,  72, 232, 104, 194,  66, 226,  98, 202,  74, 234, 106
  Data.a  48, 176,  16, 144,  56, 184,  24, 152,  50, 178,  18, 146,  58, 186,  26, 154
  Data.a 240, 112, 208,  80, 248, 120, 216,  88, 242, 114, 210,  82, 250, 122, 218,  90
  Data.a  12, 140,  44, 172,   4, 132,  36, 164,  14, 142,  46, 174,   6, 134,  38, 166
  Data.a 204,  76, 236, 108, 196,  68, 228, 100, 206,  78, 238, 110, 198,  70, 230, 102
  Data.a  60, 188,  28, 156,  52, 180,  20, 148,  62, 190,  30, 158,  54, 182,  22, 150
  Data.a 252, 124, 220,  92, 244, 116, 212,  84, 254, 126, 222,  94, 246, 118, 214,  86
  Data.a   3, 131,  35, 163,  11, 139,  43, 171,   1, 129,  33, 161,   9, 137,  41, 169
  Data.a 195,  67, 227,  99, 203,  75, 235, 107, 193,  65, 225,  97, 201,  73, 233, 105
  Data.a  51, 179,  19, 147,  59, 187,  27, 155,  49, 177,  17, 145,  57, 185,  25, 153
  Data.a 243, 115, 211,  83, 251, 123, 219,  91, 241, 113, 209,  81, 249, 121, 217,  89
  Data.a  15, 143,  47, 175,   7, 135,  39, 167,  13, 141,  45, 173,   5, 133,  37, 165
  Data.a 207,  79, 239, 111, 199,  71, 231, 103, 205,  77, 237, 109, 197,  69, 229, 101
  Data.a  63, 191,  31, 159,  55, 183,  23, 151,  61, 189,  29, 157,  53, 181,  21, 149
  Data.a 255, 127, 223,  95, 247, 119, 215,  87, 253, 125, 221,  93, 245, 117, 213,  85
  Bayer16x16_end:

EndDataSection

Posted: **Wed Jul 10, 2024 2:57 pm**

wilbert wrote: Tue Jul 09, 2024 9:13 pm
miskox wrote: Tue Jul 09, 2024 6:22 pm Wouldn't this be faster:
Code: Select all
M(x,y)=temp/d/d
No, the opposite.
A division is a lot slower as a multiplication.
If you are aiming for performance, try to avoid the division (and modulo) operator as much as possible.
In this case the M array is only 8x8 so it's not that important but if you would have a very big array, the best thing would be to calculate 1/(d*d) before the loop and multiply temp with that.

Thanks for clarification. I thought that parantheses would cause a delay...looks like I was wrong.

Thanks again.
Saso

PureBasic Forums - English

Help converting short p5.js program to Purebasic

Re: Help converting short p5.js program to Purebasic

Re: Help converting short p5.js program to Purebasic