Gameboy Development Forum

looks great!

i will play this when it comes out

(source code would be great also!)

Aha, found the bugger in string.h aand my CPU usage has dropped by ~10% on the gameboy classic version, nice!

I wander if when I've got a bit further if I can get it to reference other track data dynamically (probably via pointers I'm guessing) to turn this around.

So far the game weighs in at 13kb so there's plenty of space available if I choose not to.  But it just makes sense to do it for simpler scene management.

Got some more of the title sequence in the game now and I'm thinking about how to add some sorta Border for Super Gameboy features.

No docs in GBDK (aside from SGB detection) but I guess dipping into ASM would make this possible somehow? Any ideas would be appreciated. I've only done a lil bit of ASM on the VCS 2600 so far.

Been away from the code for a while but back on to it tonight.

Linked the intro screens into the game and did a rough colour job on them for colour gameboy stuff.

Still unsure how to tackle SGB stuff though.

Orr maybe it won't have to wait after all.

I'm attempting to get my head around how to write ASM (which can be included inside GBDK, handy that) and have the following pseudocode

// sgb sound function
disable interrupts in GBDK

SET P14 and P15 to LOW?? (Looks like a value of 20 in decimal over here?)
Where are these located in the joypad register? 0xFF00 is just one address right?

nop for a moment [15us]

01000001    // FUNCTION 8-SOUND [5] - 1 PACKET [3]
00000001    // EFFECT A - NINTENDO NOISE ![8]
00000000    // EFFECT B - NONE[8]
11010000    // A PITCH[2], A VOLUME[2], B PITCH[2], B VOL[2]
00000000    // MUSIC SCORE CODE[8]
00000000    // REST IS UNUSED
00000000
00000000
00000000
00000000
00000000
00000000
00000000
00000000
00000000
00000000    //OBLIGATORY BLANK BIT[8]

nop some more for 4 frames or ensure no other ops for 4 frames one way or another.

Can anyone shed some light on this or point me in the right direction? I'm confused as to where P14 and P15 would lie in memory, 0xFF00 is just a sinlge 8 bit value. Or am I missing something?

If I can figure out how to write something generic so it could be used inside gbdk by others that would be great, but have so far only done a lil bit of atari 2600 asm.

hi, your games as far as i checked them (space invaders, and another one), don't use halt opcodes to wait. i think they run at unlimited framerate (so 200-300 fps or so).

it's proper/recommended to, when the frame is done, wait using halt, and to then do the next frame at the vblank interrupt, and read the joypad first thing after the start of the vblank, and have the joypad values then influence the immediate next frame.

if a game does not use halt opcodes, it will use up the gameboy's batteries unnecessarily fast, and emulators will use more CPU to emulate it. you can check the "CPU usage meter" in bgb's debugger, the green/amber bar on the right. if the game uses halt opcodes, it will be partially or mostly dark. if it's solid on/bright, the game is not using halt opcodes.

The joypad hardware is just a matrix of 4x2 connections. 2 of them (P14 and P15) are outputs, the rest are inputs. The SGB uses the 2 outputs for command information because the CPU didn't have to be modified to be used this way, the 2 outputs from the CPU were already there, as well as the 4 inputs. http://bgb.bircd.org/pandocs.htm#joypadinput

Right, I think I'm getting somewhere now with my sgb code.

// INSIDE N ASM FILE - IS THIS 5 microseconds? Not sure how long NOP lasts for but I'm imagining not very long.
    .globl _paus
    _paus:
    NOP
    NOP
    NOP
    RET

// INSIDE A ROMBANK in GBDK
            //SEND A COMMAND TO THE SUPER GAMEBOY - START
            *(char*)0xFF00 = 0;        //BOTH DOWN
            paus();
            *(char*)0xFF00 = 48;    //BOTH HIGH
            paus();paus();paus();
            
            for (i=0; i!=255; i++){
            *(char*)0xFF00 = 16;    // BIT 0
            paus();
            *(char*)0xFF00 = 48;    //BOTH HIGH
            paus();paus();paus();
            }

Does this look closer to something that might work? I imagine I could remove part of the delay to copy the replacement register value over from an array in the future. A whole bunch of zeroes ain't gonna do anything.

Why not simply use the timer interrupt?

It is crazily short, 5 microseconds, followed by 15 microseconds, then looped. With a long 4 frame delay between packets.

I've not done any development of late due to having to re-organize my home office into another room but hopefully should be back on it come sunday! I might put the SGB stuff on the backburner whilst I get some more of the main game ready. Might burn this onto a cartridge when it's finished though!

5 tracks down, improved the A.I. last night and feeling pretty good about things again. Will post a video / image some time this week.

This looks good so far!