Random Numbers

[Taylorover9001]

Why does random number generation suck in QB64? No matter what tricks and randomization stuff I do, it often doesn't feel right.

[FellippeHeitor]

Random number generation will go as far as QuickBASIC's would.

[bplus]

Perhaps you could provide a few code examples where things "feel" wrong?

[bplus]

I see pretty good spread of dots over the screen, no particularly dark or light spots or sections which would suck.

With magnifying glass, I see craters and even a few canals.

Code: QB64: [Select]

1. _TITLE "RND test by bplus 2018-04-04"
2. 'QB64 version 2017 1106/82 (the day before they switched to version 1.2)
3.  
4.  
5. 'Looking for fairly even spread of dots
6.  
7.  
8. CONST xmax = 800
9. CONST ymax = 600
10. SCREEN _NEWIMAGE(xmax, ymax, 32)
11. _SCREENMOVE 360, 60
12. _DEFINE A-Z AS _FLOAT
13. screenfull = xmax * ymax
14. COLOR _RGBA32(255, 255, 255, 15)
15. FOR i = 1 TO screenfull * 10
16.     x = xmax * RND
17.     y = ymax * RND
18.     PSET (x, y)
19. NEXT
20. _TITLE "Test Done!!!!!!"
21. _DISPLAY
22. _DELAY 1
23. SLEEP
24.  
25.  

[bplus]

Here is a built in magnifying view finder, maybe you can find a man on the moon or Waldo.

Code: QB64: [Select]

1. _TITLE "RND test by bplus 2018-04-04"
2. 'QB64 version 2017 1106/82 (the day before they switched to version 1.2)
3.  
4.  
5. 'Looking for fairly even spread of dots
6.  
7.  
8. CONST xmax = 800
9. CONST ymax = 600
10. SCREEN _NEWIMAGE(xmax, ymax, 32)
11. _SCREENMOVE 360, 60
12. _DEFINE A-Z AS _FLOAT
13. screenfull = xmax * ymax
14. COLOR _RGBA32(255, 255, 255, 15)
15. DIM magnify(xmax, ymax)
16. FOR i = 1 TO screenfull * 10
17.     x = xmax * RND
18.     y = ymax * RND
19.     PSET (x, y)
20.     magnify(x, y) = magnify(x, y) + 1
21. NEXT
22. _TITLE "Sceen Test Done"
23. _DISPLAY
24. _DELAY 1
25.  
26. 'make sure my viewer is aligned right by using a black line in the middle of screen
27. FOR i = -15 TO 15
28.     magnify(xmax / 2, ymax / 2 + i) = 0
29. NEXT
30.  
31. LOCATE 15, 10
32. INPUT " OK press enter to see mouse over x, y  40 x 30 block magnfied."; magnify$
33.  
34. DO
35.     CLS
36.     FOR y = 0 TO 600
37.         FOR x = 0 TO 800
38.             COLOR _RGB32(magnify(x, y) * 15, magnify(x, y) * 15, magnify(x, y) * 15)
39.             PSET (x, y)
40.         NEXT
41.     NEXT
42.  
43.     DO WHILE _MOUSEINPUT
44.         mx = _MOUSEX
45.         my = _MOUSEY
46.     LOOP
47.  
48.     IF mx > 20 AND mx < 780 AND my > 15 AND my < 585 THEN
49.         COLOR _RGB32(0, 0, 0)
50.         LINE (mx - 20, my - 15)-(mx + 20, my + 15), , B
51.         FOR y = -15 TO 15
52.             FOR x = -20 TO 20
53.                 COLOR _RGB32(magnify(mx + x, my + y) * 15, magnify(mx + x, my + y) * 15, magnify(mx + x, my + y) * 15)
54.                 LINE (5 * (x + 20), 5 * (y + 15))-STEP(5, 5), , BF
55.             NEXT
56.         NEXT
57.     END IF
58.     _TITLE "Mouse x" + STR$(mx) + ", Mouse y" + STR$(my) + "       press spacebar to quit."
59.     _DISPLAY
60.     _LIMIT 60
61. LOOP UNTIL INKEY$ = " "
62.  
63.  

[Bert22306]

Yeah, I also wrote a "dots on screen" test, plus a histogram test to show occurrences in 10 separate bins, as well as min, max, and mean, and these tests look good. Nothing obviously bad about the built-in PRNG, even if the seed is only something like 24 bits wide.

Why do you think it sucks?

[FellippeHeitor]

Here is a built in magnifying view finder, maybe you can find a man on the moon or Waldo.

That was neat bplus!