This is mostly just the implementation of a fundamental Game Boy emulator, but there are some differences that I might like to speak (or write) about. I am referring to the reference implementation of the Raspberry Pi Pico and the Raspberry Pi Foundation’s minimal hardware instance, which run at 125 MHz. 2. We lose the reference clock when the sport Boy enters the halt state, which for most video games happens at the very least once per body. Since we get the exact clock from the sport Boy we will depend an emulated register in synch with the true one without any danger to diverge. So, possibly we will do a fractional clock rely, however for now as it solely affects the div register which I cannot properly initialize anyway, this is not carried out. Here we can observe hundreds of cycles and will be able to get a very precise substitute clock from our rp2040. This also goes for the choice whether or not a block of the garbage pile in mode B is empty or filled, so we also get the identical layout for the rubbish stack. So, the game Boy that starts the game first will generate that stack and ship it to the second by way of link cable.
Oh, and while interrupts are a blessing to synch the PPU, these aren’t exactly straightforward to detect in the primary place. First of all, there are a number of things that change into a lot less complicated on this scenario. Since we don’t want to switch the video games, there will not be a lot of curiosity on that port. While there are various site with opcode tables for the sport Boy, I found the one by Megan Sullivan to be the most handy one, which is why I had it open pretty much all the time. However, there will probably be video games out there with a different approach (like my Wifi cartridge) and the output of the GB Interceptor will look glitchy until a detection for these other strategies has been applied. Most games simply let the sport Boy trigger an interrupt when vsync is reached and we can see when the code of this interrupt is being executed, so we will simply regulate the timing of our personal emulated PPU to enter vsync at that same moment. The massive drawback here is that games can use many alternative ways to do so. The issue is, that no hint of the PPU will be found on the reminiscence bus.
This solely becomes an issue if the info from the I/O registers finally ends up in VRAM. We must deduce the state of the PPU by the habits of the sport, which has to synch to the PPU as properly – a minimum of to know when it might write to VRAM. This result’s then written to VRAM and we don’t know what is in that location. The second writes the data on to VRAM with none checks or conditional jumps and we can’t see something. In particular, we do not see the input from the sport pad! The true Game Boy will fetch the following instruction and we don’t need to care about the place the instruction came from. Our emulator will simply comply with these same directions and doesn’t have to care whether or not it was triggered by a button press. Well, nearly each code in existence will compare the gamepad enter to examine which button was pressed and make a conditional soar to code that’s triggered by the button. Imagine that the worth of the gamepad is added to a base address to calculate the tile index to an image that reveals the current state of the D-Pad.
The CPU would get the instruction to fetch the gamepad register worth, add a quantity to it and our emulator would not know the correct result of that operation. We only get into real trouble when an entire stream of prepared information is written to VRAM from one of many I/O registers. The second core then acts as the PPU and renders a picture from our VRAM copy. Otherwise any effect that requires adjustments in the VRAM midframe would lead to glitches and at the least we’d see some tearing effects as information is updated in VRAM randomly. Therefore, it’s not shocking that we will see a few differences right here between different units. Here, we are able to look at the same example of the mode B rubbish stack, however in two player mode of Tetris. Many hardware details and a few intricate specifics might be found on Joonas Javanainen’s webpage. This solves one of many seemingly greatest points: We can’t see any of the hardware I/O registers. How ought to we ever emulate a sport if we can’t see the player’s enter?
