Prizm Useful Routines -- post here!

[Ashbad]

Since Cemetech has one, I think we need one as well.  This section is where you can post any routines for the casio Prizm you make in C or SH4 assembly.  I don't think I have the liberty to post others' routines over here, but I'll be glad to start with some of my own, and I'm sure Qwerty will soon post some of his very well done SH4 routines [nice and optimized, too!] such as his LRNG.  Here are a few in C, all written by me, and decently optimized:



Routine: Draw Alpha-coded sprite

inputs: pointer to data, x pos, y pos, width, height, alpha degree 1-255 (0 denotes no alpha value and normal drawing)

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8 9 10 11 12

void AlphaSprite(short* data, int x, int y, int width, int height, char alpha) {    short* VRAM = (short*)0xA8000000;    int CurColor = 0;    VRAM += (LCD_WIDTH_PX*y)+x;    for(int j=y; j<y+height; j++) {       for(int i=x; i<x+width;  i++) {         CurColor = (*(VRAM) + (alpha*(*(data++))/256))/2;         *(VRAM++) = CurColor % (65536);       }       VRAM += (LCD_WIDTH_PX-width);    } }

Generate pseudo-random number

input - 32bit integer for upper and lower bounds

this is based on the famous '128bit shifting' technique, and is considered highly random enough.

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8 9 10 11 12 13

int Rand32(int lower, int upper) {   static int a = 123456789;   static int b = 362436069;   static int c = 521288629;   static int d = 88675123;   int t;     t = a ^ (a << 11);   a = b;   b = c;   c = d;   return (d = d ^ (d >> 19) ^ (t ^ (t >> 8)))%upper + lower; }

Test a pixel of CURRENTLY rendered VRAM image (tari and Ashbad optimized)

inputs: X and Y position
outputs: short holding color value

Code: Select All | Copy To Clipboard

1 2 3 4

short getpixel(short x, short y) {     short *VRAM = (short *)0xA8000000;     return *(VRAM + (y * LCD_WIDTH_PX) + x); }

Invert Colors in Rectangular Area

inputs: shorts for x position, y position, length, and width
outputs: look at your screen!  (if you actually render it)

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8

void InvArea(short x, short y, short height, short width)    unsigned short *VRAM = (unsigned short *)0xA8000000;    for(short a = 1; a>width; a++) {       for(short b = 1; b>height; b++) {          *(b + y * LCD_WIDTH_PX + a + x + (VRAM++)) ^= 0xFFFF;       }    } }

Print Custom Character (improved)

inputs: an x position, a y position, a bit mapped image of the character, color to draw the character, if it is to draw a backcolor in non-mapped spaces, a short to specify backcolor (put 0 if DrawBackColor is false) , and the dimensions of the character
  EXAMPLE BIT MAPPED CHARACTER (the letter A):
          [0,0,1,0,0]
          [0,1,0,1,0]
          [0,1,1,1,0]
          [0,1,0,1,0]
          [0,1,0,1,0]
outputs: look at your screen (unless you don't render it)
       
       
       

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8 9 10 11 12

void PutCustC(*bool map, short x, short y, short width, short height, short color, bool DrawBackColor, short backcolor) {      short* VRAM = (short*)0xA8000000;      VRAM += (LCD_WIDTH_PX * y) + x;      for(short a = 0; a>width; a++) {                for(short b = 0; b>height; b++) {                          if(*(y + b * width + x + a + map)) { *(VRAM++) = color; }                          elseif(DrawWhite) { *(VRAM++) = backcolor; }                          else { VRAM++; }                }                VRAM += (LCD_WIDTH_PX-width);      } }

Draw Circle Routine

inputs: x, y, draw color and radius
outputs: look at your screen (that is, if you render it)

NOTE: uses a SetPoint routine in display_syscalls.h

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

void DrawCircle(short x0, short y0, short radius, short color) {   short f = 1 - radius;   short ddF_x = 1;   short ddF_y = -2 * radius;   short x = 0;   short y = radius;     Bdisp_SetPoint_VRAM(x0, y0 + radius, color);   Bdisp_SetPoint_VRAM(x0, y0 - radius, color);   Bdisp_SetPoint_VRAM(x0 + radius, y0, color);   Bdisp_SetPoint_VRAM(x0 - radius, y0, color);     while(x < y)   {     // ddF_x == 2 * x + 1;     // ddF_y == -2 * y;     // f == x*x + y*y - radius*radius + 2*x - y + 1;     if(f >= 0)     {       y--;       ddF_y += 2;       f += ddF_y;     }     x++;     ddF_x += 2;     f += ddF_x;        Bdisp_SetPoint_VRAM(x0 + x, y0 + y, color);     Bdisp_SetPoint_VRAM(x0 - x, y0 + y, color);     Bdisp_SetPoint_VRAM(x0 + x, y0 - y, color);     Bdisp_SetPoint_VRAM(x0 - x, y0 - y, color);     Bdisp_SetPoint_VRAM(x0 + y, y0 + x, color);     Bdisp_SetPoint_VRAM(x0 - y, y0 + x, color);     Bdisp_SetPoint_VRAM(x0 + y, y0 - x, color);     Bdisp_SetPoint_VRAM(x0 - y, y0 - x, color);   } }

EDIT: and I'm not sure where to put this, but here's a small yet useful update on the BFILE_syscalls.h header, with defines that more easily stand in place of a number for 'mode' in the Bfile_OpenFile_OS function.

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8 9 10 11 12 13 14

#define O_READ 0x01 #define O_READ_SHARE 0x80 #define O_WRITE 0x02 #define O_READWRITE 0x03 #define O_READWRITE_SHARE 0x83 int Bfile_OpenFile_OS( const unsigned short*filename, int mode ); int Bfile_CloseFile_OS( int HANDLE ); int Bfile_ReadFile_OS( int HANDLE, void *buf, int size, int readpos ); int Bfile_CreateEntry_OS( const unsigned short*filename, int mode, int*size ); int Bfile_WriteFile_OS( int HANDLE, const void *buf, int size ); int Bfile_DeleteEntry( const unsigned short *filename ); void Bfile_StrToName_ncpy( unsigned short*dest, const unsigned char*source, int n );

[fxdev]

Please, use Syscall 0x01E6 to obtain the VRAM address.
Interface: void *GetVRAMAddress(void);

[DJ Omnimaga]

Thanks Ashbad for those routines. Hopefully they might be useful for some people here too. Some other people should post routines here too if they have any.

[Ashbad]

Please, use Syscall 0x01E6 to obtain the VRAM address.
Interface: void *GetVRAMAddress(void);

I could, but since the address is known, it's a whole lot easier in many cases just to remember the magic number 0xA8000000 however, good point brought up.

I just updated the first post at the bottom with an updated BFILE_syscalls.h header, which will prove to be very useful if you're opening any files at all.

[DJ Omnimaga]

Could the address possibly change in future OS versions? I hope not....

[Ashbad]

Could the address possibly change in future OS versions? I hope not....

From what I know, I don't think it would, then many other things would also get moved around, so I think for now we're safe

[AngelFish]

I see no reason why they couldn't change it, to be honest. All it would take is a change in the TLB code to move it. A few hundred bytes right or left isn't that much stuff to rearrange. They did that to make the Prizm OS, if anyone remembers.

[DJ Omnimaga]

Yeah true. I just hope it won't change too much from an OS to another. Remember the mess when OS 2.53MP got released

[Ashbad]

a few new routines, obviously ripped from somebody else and not by me, but however, I shall give full credit to the genius author of these awesome routines, prizmized by me.  EDIT: I was not able to easily convert a generic polygon-filling C routine to a decent Prizm-like format, so feel free to find/make your own

See if point is within bounds of a polygon

inputs:
int    polySides  =  how many corners the polygon has
float  polyX[]    =  horizontal coordinates of corners
float  polyY[]    =  vertical coordinates of corners
float  x, y       =  point to be tested

outputs:
returns if the point is within the bounds of the polygon.

Credit: many thanks to Darel Rex Finley, http://alienryderflex.com/polygon/

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8 9 10 11 12 13 14

bool pointInPolygon(int polySides, float polyX[], float polyY[], float x, float y) {   int      i, j=polySides-1 ;   boolean  oddNodes=0      ;   for (i=0; i<polySides; i++) {     if (polyY[i]<y && polyY[j]>=y     ||  polyY[j]<y && polyY[i]>=y) {       if (polyX[i]+(y-polyY[i])/(polyY[j]-polyY[i])*(polyX[j]-polyX[i])<x) {         oddNodes=!oddNodes; }}     j=i; }   return oddNodes; }

[fxdev]

I could, but since the address is known, it's a whole lot easier in many cases just to remember the magic number 0xA8000000.

Could the address possibly change in future OS versions? I hope not....

They did this with OS 2.00 for the fx-9860G/GII back in 2009 and broke several hard-coded add-ins. So, do not rely on 0xA8000000.

[DJ Omnimaga]

Ouch this sucks. I think it would be best to avoid direct address usage then. I think this is what we do on some 83+ stuff.

[z80man]

Please, use Syscall 0x01E6 to obtain the VRAM address.
Interface: void *GetVRAMAddress(void);

Due to my dislike of any syscalls I will just write a routine that does the exact same thing and polls the same address from ROM. It sounds foolish, but that is just the way I roll 

Note: I will go as far as file systems and text routines when it comes to avoiding syscalls.

[fxdev]

Okay, your decision but as long as a Prizm routine isn't buggy and fits my needs, I won't rewrite the whole thing.

[z80man]

Alright I have a conversion from C to asm. It is get pixel and is as fast as possible and can be no more optimized. Unless gcc automatically handles the zero extension which I'll have to check.


Test a pixel of CURRENTLY rendered VRAM image (z80man optimized)

Code: Select All | Copy To Clipboard

1 2 3 4 5 6 7 8 9 10 11 12 13 14

short GetPixel(short x, short y) {     MOV.W (width),R2     MOV.L (VRAM),R3     MULU.W r2,R5     ADD R4,R3     STS MACL,R2     ADD R3,R2     MOV.W @R2,R0     RTS     EXTU.W R0     align.4     width: word 384     VRAM: long VRAM_address

[Ashbad]

this is awesome I would suggest putting it into .s format so C coders can use it.