moa/todo.txt
transistor 941c793523 Added pixel encoding, requested by frontend
The hope was that this would reduce the amount of copying and bit
shifting required by the frontend to get the data on screen, but
it doesn't seem to offer much advantage, surprisingly.  I'll leave
it in though.  There are a few other minor tweaks included here to
try to improve the performance a bit
2023-03-25 10:51:46 -07:00

131 lines
7.1 KiB
Plaintext

* improve performance
* should it be possible to reschedule multiple events at different intervals to reduce the times a given step function is called? Some have
multiple clocks, or multiple things at different clocks, and making them each an event would mean they could be smaller and faster, but at
the cost of having more events on the queue when re-scheduling. There needs to be a mechanism to avoid the event queue ballooning due to
an error
* can you refactor the update timeout to put it in rust? Would that make it faster? (the tricky part is the closure)
* re-enable sound on webassembly, see if it works (it does not. Very lagged and jittery with what sounds like repeated frames)
* can you somehow speed up the memory accessing through the sim? The dyn Addressable is causing a fair amount of overhead
* can you somehow make devices have two step functions for running things at different times? (I'm thinking ym2612 audio gen vs timers)
* should you rename devices.rs traits.rs?
* you could refactor the instruction loop into a series of functions, and test if there's a performance difference with and without #[inline(always)]
* move parser into its own utils/ or /libs/ directory
* fix ym2612 sound generation
* need to re-add a mechanism for audio frame dialation, either based on speed, or somehow automatic
* sound doesn't work on a lot of games... is it a problem with the Z80 accessing the YM2612, or the lack of YM timers? or something else?
* the pixel format idea didn't work because of window resizing and the fact that the frame needs to be adjusted in size because the window can't always be resized...
* add mouse support to synth app
* test the Z80 more, add tests
* double check the functioning of the banked areas and register settings for Z80 coprocessor
* make the keys easier to config...
* can you make the debugger more accessible, so a web interface could access the data and display it, in light of the fact that println isn't available in wasm
Web Assembly:
* can you make the web interface nicer with like... a picture of a genesis or something
* would a different pixel format help at all?
* can you limit the size of the window that pixels generates?
* can you automatically adjust the speed based on the calculated framerate (if you moved that to Rust)
* enable sound in web assembly
* add ability to disable one or the other audio chips in the genesis
* make it possible to disable audio in browser
* make it possible to compile without audio support (minifb frontend requires it atm)
* should you have a separate attenuation value for each input in the mixer so that you can make one chip quieter (the sn76489 is pretty loud, and I added a fixed offset to the attenuation for now)
Audio:
* address repeater on ym2612 doesn't seem to work the same, when it's on the 68000 device. The Z80 device doesn't have an affect, but maybe it's not being used
* should you represent audio as frequencies rather than amplitude so that time dilation is more accurate? Would possible require less
accurate simulation of the audio
* Removing at a sample-level granularity would compress or lengthen the waveforms, so it would be better to mix/drop a whole chunk at
once (either predetermined by the audio system or determined by each device by the amount of samples it writes at once). The chunk
size could either be specified by the device in microseconds or something, or can be inferred by the sample_rate and the size of the
chunk.
* you could make the sound device be an object that is passed back to the simulation section like SimplePty. You need to either register
a callback with the frontend sound system that is called when it needs data, or you write to a shared buffer which is passed back to the
frontend when it needs it, or it has a copy it can use directly
System/Traits:
* can you make the connections between things (like memory adapters), be expressed in a way that's more similar to the electrical design?
like specifying that address pins 10-7 should be ignored/unconnected, pin 11 will connect to "chip select", etc
* should you add a unique ID to devices, such that they can be indexed, and their step functions can reset the next_run count and run them immediately
* should you simulate bus arbitration?
* interrupts could be done in a better way
* need a better way of handling disparate reads/writes to I/O spaces, rather than having multiple devices or having a massive chunk of address space allocated, continuously
* should you modify Addressable to also take the absolute address as input? I'm thinking of how the same device could be mapped to multiple addresses in memory instead
of taking up a whole range of addresses
* you could modify read()/write() in Addressable to return the number of bytes read or written for dynamic bus sizing used by the MC68020+
Debugger:
* i need a way to debug only the cpu and not the coprocessor, but that's tricky without a way to id or compare Transmutables
* add a way to delete a watcher
* can you improve how the watcher implementation in the Bus works, instead of setting a flag and then checking it every cycle, pass in the System to Addressable??
* can you use the breakpoint address parser in other commands?
* get stack tracing working again, but can you do it with just data?
* how can you improve the debugger?
* the command line definitely needs to be fixed so it prints the prompt correctly
* debugger could maybe even allows arrows left/right for editing, and up/down for history
Genesis/Mega Drive:
* there is an issue with Mortal Kombat 2 where it will crash randomly at the start of a fight. The code is actually swapping
stacks a bunch of times, and at some point, the stack is corrupted or something and it `rts`s to the wrong address...
* implement sn76489 and ym2612 for audio
* there's a bug when Sonic 2 goes to the demo screen, it's all corrupted (could it be a dma copy error)
* the 68000/Z80 bank switching is probably buggy
* the H/V counters are not accurate because it seems to count at different speeds in the blanking period (time vs return value numbers don't divide properly)
* make the ym7101 set/reset the v_int occurred flag based on the interrupt controller
* add support for the sprite overflow flag (low priority)
* still possibly a bug with the shadow/highlight colours
Macintosh:
* issues when booting the rom, attempt to write to rom during the driver init/open phase
* for the address bus/repeating thing in the mac with the rom and ram, can you make it work for both the 128 and 512
68000:
* unimplemented: BFFFO, BFINS, NBCD, RTD
* >=MC68020 undecoded & unimplemented: BKPT, CALLM, CAS, CAS2, CHK2, CMP2, RTM, PACK, TRAPcc, UNPK
* add support for MMU
* add support for FPU
* Coprocessor instructions: cpBcc, cpDBcc, cpGEN, cpScc, cpTRAPcc
Z80:
* add instruction timings to Z80
* unimplemented: CPD, CPDR, CPI, CPIR, DAA, IND, INDR, INI, INIR, INic, INx, OTDR, OTIR, OUTD, OUTI, OUTic, OUTx, RETI, RETN, RLD, RRD
* can you eventually make the system connections all configurable via a config file?