#Super Nintendo/Famicom
1 messages · Page 11 of 1
thats very cool. hopefully srg and you can get them working on mister 🙂
I don’t have a MiSTer FPGA, but I would be really happy if you could play it and hold tournaments like the ones in 1992 and 1994.
I also need people who have an FXPAK PRO or SD2SNES to run tests and ensure these games work perfectly.
Guaranteed there’s a few people in this Discord with them
I have one in my super nt box
Can you tag those people? I want to start right now, since I’m on vacation—I’m a university and primary school teacher.
Can you share screenshots of your Super NT box?
Does the version matter? I think my sd2snes is the non pro
I sent you a private message.
Please tell the people who have an FXPAK PRO or SD2SNES to send me a message.
Isn't that just any emulation device on Aliexpress or Temu?
Who here has these flash cartridges to get Nintendo Campus Challenge and PowerFest ’94 working?
https://krikzz.com/forum/index.php?board=18.0 this is the support forum for those
making an account there and requesting support would probably be the best path
but i'm not sure if they work or not
there are quite a few people with those carts here
I take it back, my sd2snes is not near the super nt. which begs the question "where is my sd2snes". But yeah, as birdy mentioned, probably best to go to the source
i have one but my SNES is dead X_X
i necro'd the forum post there - https://krikzz.com/forum/index.php?topic=8764
Competition carts on SD2SNES Pro?
see if we can get the convo going
I both don't have a SNES flashcart and the sound stopped working on mine 🙁
I have a standard SD2SNES
people with no life
It's a dolphin! 
Ok, honestly, I tried and normally am good at these but the awkward shape kills it for me
it don't think it's anything. looks like a 7 that someone got tired of drawing half way through
I have a sd2snes pro/fxpak pro and a USA snes and a super nt , happy to help test stuff
- unstable changes from srg320 on 2025-11-18 (PPU: fix sprite evaluation during Force Blank)
- PPU: Fixes for 16x32 & 32x64 sprites and tile count/flip commit from paulb-nl
- ACLK_FREQ to 410496 (fixes the title screen on Unholy Night) - DSP_PKG.vhd
- latest sys module changes from Taki
damn, so hacks using SA-1 in this configuration won't run on mister? as someone who loves SMW hacks, that's a shame
wasn't there an alternate SNES core that mimics this non-standard behavior? i don't recall
So many cores across systems being updated because of a framework change, what was changed in the framework?
Nothing super important. Taki must need it for something
I can see who that is but the helmet also kinda looks a little bit like Serpentor's helmet a teensie bit too
yes there is
SNESa1 for MiSTer. Contribute to littlegamer87/SNESa1_MiSTer development by creating an account on GitHub.
thank you! this is really cool, i'm glad it exists
i wonder if we'll see this get merged into the main core eventually?
well paul doesn't oppose it anymore, on the condition that it is implemented as an option and disabled by default
srg wanted to update the sa1 for accuracy because now it runs too fast and wanted to keep some logic headroom
if there is enough headroom left after his updates I can do a PR for this
pretty sure the CX4 code also has a patch for rom hacks that make the same mistake, I can lock that behind an option as well
neat, that would be awesome. would be good for it to be disabled by default, but more options are always welcome when they open up access to more games/hacks
interested to see what srg has in store for the core as far as fixes go. i was under the impression that the SA-1 implementation here was already better than what the FXpak pro has
could be, hard to compare
fxpak does allow rom hacks to do the same stuff as this fork, so mister is the only platform where these hacks wont run
I don't get it. Will every SA-1 romhack stop to work? Or just some poorly implemented ones?
there are a very few that will run. most don't
the ones that run have been fixed in revision 1.04 in the hack or something
kinda chicken and the egg story because most development platforms wont warn the rom hacker that there is something wrong, unless you use a real cart with a maskrom replacement, or have access to mister
1.04 of which hack? (Sorry, I should probably just read up)
take megaman x sa1 as an example
So 1.04 will work on real hardware and MiSTer, 1.03 will not
1.02 in this case, the older ones also didnt work. it depends how much maintenance and testing the hacker did or had access to
I am honestly all for those being added as options disabled by default, I have seen similar options to enable inaccurate behavior in newer emulators for the sake of playing older romhacks that relies on said behavior. As long as they are not on by default and do not compromise actual accuracy just by being included, I think it's a good idea.
Yes, all default options in MiSTer cores must be as close to the original as possible.
That in of itself is a form of preservation; stock experience is how the console was.
This is one of the best things about MiSTer imo
gonna dip you in resin and preserve you
PPU 2 🙂
Just tried that test version but the change in audio clock frequency made Bad Apple demo worse for me. The sound cuts earlier than it did previously and i also got a few sync audio glitches at the beginning of the demo. It would be nice if the ACLK_FREQ could be toggled in the core menu 🤔
We will see later if that's possible.
the point of debate here is patching the sa1 to get these rom hacks to run won't change the behavior of the published sa1 games at all
in latent's terms, let me make a comparison by calling it tolerance
the original sa1 had a tolerance of 5mm (sorry americans), so the published games were designed to work with 5mm tolerance in mind
emulators back then knew what the sa1 was doing but didnt have a clue on the tolerance, that knowledge came way later once the dev manuals were leaked. on an emulator, you could easily do 1mm tolerance
the published games required 5mm and would run with no issue if it was tighter, still within range
these rom hackers also didnt have a clue about it either so they did not take the 5mm tolerance into account, they just thought it was 1mm
and then suddenly the mister core comes along that actually knew what the tolerance was, and so these rom hacks are broken
so if you would make the core's tolerance also 1mm, does that conflict with preservation? that's debateable I guess
I get that, I just feel that if something is broken or has bugs/glitches on what would basically be real hardware (like replacing the rom in a SA1 cart like you said) that it would be more accurate to have the mister core act that way too, at least by default
People might use the mister core to develop homebrew meant to run on real hardware, you would want the core to be as accurate as possible for that
yes, but on real hw, replacing maskrom on a real sa1 cart is not plausible, that's almost vandalism 😄
the only plausible way is flash carts, and the ones that can do sa1 are very limited. think krikkz is the only one that supplies them
I guess krikkz realized the same thing - setting it tighter does not break anything, so then everybody thinks, runs fine on real hw
For now at least, I mean, back when our only real choices were ZSNES and SNES9X, of which most chose ZSNES, there were romhacks made on that which break on more accurate emulators.
There might be other romcarts in the future that are more accurate, or an easier way to implement an SA1 in a homebrew cart
And you never know if some homebrew performs something that somehow breaks if the tolerance is tighter for whatever reason
Some homebrew in the future I mean
I need the help of this community. I have a 1 TB Samsung microSD card in my MiSTer. At some point after an all-update, I noticed that the SNES core started taking significantly longer to load games. All other cores run without any issues. What could be the cause of this?
Yes, the core needs to work exactly like original hardware, if those rom hacks don’t work with original hardware then they need to be fixed.
Software emulators should never have been adjusted to support broken rom hacks or rom dumps imo.
The problem is that people don’t want their flash carts or emulators to appear like they don’t work so it creates this cycle of continually supporting broken roms and rom hacks.
People need to be educated that if a rom hack doesn’t work then it’s the fault of the rom hack and not the emulator or flash cart. Like full stop, that has to be the assumption moving forward.
I don't think this is a good analogy. It's not an issue of "tolerance" at all, it's an actual difference in behavior between the original hardware and a bad emulation of it. "Preservation" also means preserving the original hardware, not just the games.
Also, like I don’t want recreation of original hardware to be compromised just so I can play the “Mario takes a dump at the Zoo” rom hack.
In many cases (I don't know about this one) differences in behavior that usually don't matter for normal gameplay often become very important during speedruns or other contexts involving heavy glitches
and probably the most important point: the reason why we have crappy romhacks that don't work on original hardware is because we had bad emulators that didn't emulate the original hardware properly
if we continue to make bad emulators on purpose, then the result of that is more romhack authors making hacks that don't work on original hardware
Do you have a link to that Rom hack sir?
The more accurate MiSTer is, edge cases and all, the more useful it is to anyone making things (as opposed to just playing games)
To be fair that is an exquisite romhack, though not as good as Yoshi vomits in a Ballpit
that's a fair point, I did not consider that
lol, this sounds like you are referring to mario's mystery meat 😄 that is an sa1 hack impacted by what I mentioned, but it's actually considered a classic in the mario community
slow loading times are very likely to be an sd card issue. try another one, or benchmark it to see if it can do the speeds it promises (most effective way to identify a fake)
If it were caused by the SD card, the long loading time would affect all cores. However, it only affects the SNES core.
The screen stays black for a few seconds until the ROM’s loading screen eventually appears.
no, it is not as black and white. existing data may be fine but newer data written to it can cause degradations
i recommend to only rule it out once tested
Thanks for the hint! I’ll try it out right away.
lmao that’s a funnier one than mine
i do largely agree with this, but i think having a toggle for SA-1 behavior would be a win-win for everybody. leave it set to "accurate" by default, and for those of us who want to play the affected ROM hacks, have an "inaccurate" option. you wouldn't have to compromise anything
obviously the mission statement of the mister is to be as close to a 1:1 replication of the original hardware as possible, but i don't think there's anything wrong with having extra bells and whistles on top of that
I have no idea what you are talking about, I would know absoultely nothing about a romhack based on Vinny Vinesauce memes or how it's a reference to his band Red Vox's song Vomit in the Ballpit. You must be imagining things. 😛
Also the romhack was updated shortly after release so it works properly on just about anything IIRC. Before it would only work on a specific slightly older version of SNES9X
Agree unless those extra bells and whistles eventually make it harder to hit the 1:1 with original hardware goal
Hmm, seems I spoke too soon. I tried playing it just now and while the usual visual glitches I have seen on more accurate emulators that happen with earlier versions of the hack from making the 5-headed-duck die or god's pool of piss worked fine, I keep running into a new glitch I had never seen before. Any time there is a dialog box the background just gets replaced with a solid white. 🙁
🎩
🧿👄🧿
✨
🫴
Oh, also I can't see the cursor in menus 🙁
mario's mystery meat you mean?
exactly. in this case, i really doubt a toggle for SA-1 behavior would take up a meaningful amount of space in the core
Yes
worth trying the alternate core?
Which one?
indeed, it either expects multiplier results in 1 cycle or uses the cpu to write to sa1 registries, not sure which one it is. runs fine on my snesa1 fork
#1096636309679919136 message
I am guessing I would not be able to load a savestate from the stable core in that one?
might be worth asking @vast oak to pin this core for the time being?
should work fine in general
unless the state is already broken from missing sa1 features
Oh, the stable core does not have savestates yet anyway, nevermind that then XD
Let's see if the alt core works with it then
ROBBY NO LIKEY SMART IDEAS AND COMPRIMISE, ROBBY SMASH!!!!
Ok, the alt core seemed to work. On top of the other things I mentioned working, I was no longer getting a white background from dialog boxes and I could see the cursor in menues. I didn't play through the whole thing though, but I did go through the Earthbound section, spaghetti hell, Chatyot village, SHMUP section, vaporwave area, rizon fortress and first part of will smith's rewind time with no issues. I didn't even bother attempting the verizon speed test though.
nice, glad to hear that sorted things out
happy gaming 😄
Yeah, for sure! Compatibility options are great as long as they're disabled by default and don't add a big maintenance burden.
New avatar pic, very nice
But there are implementation reasons why that can be complex sometimes (not saying those apply to this specific case, just in general):
- Options add more code paths that needs to be tested; if you have 10 toggles you have over a thousand different combinations of options that might interact with each other in different ways.
- MiSTer is a hardware project, not a software project. You can't just add a "if this option is set, do this" branch; oftentimes implementing something in a way that behaves just slightly differently involves making a hardware block that works completely differently under the hood.
A lot of software emulators (especially older, inaccurate ones) take a high-level approach to emulation: look at the documentation for some hardware feature, and write a function that implements that spec. This works for games that are coded to the spec, but can result in different behavior for games that violate the hardware specifications (like a lot of romhacks tested in these early emulators)
MiSTer takes a lower-level approach: figure out how the hardware actually works, and recreate it. All of the hardware's quirks and weird edge cases exist because of the way the original hardware was designed and implemented. So if you build a recreation that works in the same way, you naturally get the same edge case behaviors.
E.g. the CPU multiplier works by doing one shift and add per cycle, so it takes 8 cycles to complete an 8-bit multiplication. Some early emulators just did a simple multiplication in software, which completes instantly. So there were early ROMhacks that assumed the result of a multiplication would be available instantly, and so those hacks don't work on real hardware.
But since MiSTer is actually trying to recreate the hardware, the core developers will design a multiplier that works the same way, by doing one shift-and-add per cycle. Adding a compatibility option for a single-cycle multiply wouldn't be a simple if statement, it would mean designing a completely different multiplier that can complete in one cycle (and in fact that multiplier would be much larger and more complicated, which is why the original CPU doesn't have a single-cycle multiplier in the first place).
“BuT iT’s JuSt a ToGgLe
”
theres's also an argument to be made that a romhack that only works in an emulator isn't really a SNES game, it's just a PC game with extra steps
if you care about preserving the Authentic Experience of playing those romhacks, you shouldn't be going to FPGA developers, you should be going to VM developers so that you can emulate zsnes on windows XP
(that's half a joke, I get that MiSTer is a much more pleasant experience for playing those hacks, so I agree that compatibility options are a good idea but only where that's practical and pragmatic)
i'll have to disagree with that in this case. most of the people making SMW ROM hacks in the last few years had no idea their hacks wouldn't work on real hardware, since they already worked on the SD2SNES
then the right solution is for SD2SNES to be improved to match the original hardware, so that people stop doing that
It seems to be a constant cycle on here. A core is an almost perfect replication of original hardware, people find a romhack or obscure mapper for a bootleg game, then they come here complaining that the one obscure thing isn’t playable on MiSTer
nada, weren't you the one who was working on making trackers work with the mister core?
I am more excited for that tbh
yeah!
then the SD2SNES is to blame in this case for not having accurate SA-1 implementation
but my philosophy is that if it works on hardware, then it should work on mister
Problem is there’s extra hardware in the chain there 😅
yeah, again the SD2SNES is arguably at fault in this case
but it's not really the ROM hackers fault for not realizing that the emulators/flash carts are flawed
Everyone should just use a MiSTer 
ideally yes
Problem solved
I've pushed out another test build just this week fixing some stability / correctness bugs; pretty much the only thing left to do is to implement SRAM support, do some more testing, and some cleanup to get it ready for upstreaming
I'm hoping it'll be ready for review by ~the end of the year
Not just here -- this debate has happened literally thousands of times over the past 25 years in every emulator community
the consensus is always that we should have emulators that are as accurate as possible, because even though it breaks some romhacks today, it ensures tomorrow's romhacks won't have these compatibility issues
but then you have idiots like me who still enjoy older hacks 🙂
Yeah, so compatibility options are nice for preserving historically significant hacks (or you can always tell people to use an older emulator)
Free yourself of romhacks friend, the library is big enough without them 😉
yeah, but super metroid map randos are so fun
but MiSTer in particular is a project that especially goes for accuracy every time, as far as I'm aware
Pass. I generally don’t touch romhacks unless they’re translations 🙂
My backlog is big enough without them 😅
that's definitely the mission statement yes, but then some cores like the N64 core definitely cut a lot of corners to fit on the DE-10. even then, one of the main reasons people enjoy it as much as they do is because of the enhancements, like the turbo core
If you're just using basic autosplitters / trackers, what I have now should work already for you. More complex stuff like Archipelago needs SRAM support
there's also some issue with codetracker on ALTTP, but I can't really debug that right now since codetracker doesn't work on linux
rom hacks are my jam. the original library is great and all but already played the stuff i wanted to, rom hacks keep it alive
and the rom hack content for snes is huuuuuuuuuuuge
yeah, hacks and patches are arguably the main reason the SNES is one of my favorite consoles. there's just so much user-generated content out there
anyway, i still think a toggle for SA-1 behavior would be a great addition to the main core. it already exists as a separate core, so why not merge it with the main? after all, we already have options for things not possible on original hardware, like CPU overclocking, savestates, dither blending (for kirby's dream land 3), etc.
Eh, I would say that's REALLY stretching it. Also since it works in Everdrives that's also like saying you aren't real SNES games on a SD2SNES
I mean, you're asking them to release an update that will BREAK tons of romhacks that currently work and offer essentially nothing in return. I highly doubt most people are going to accept that their favorite romhack has stopped working on their expensive flashcart in the name of accuracy.
If your ROMs are zipped they probably take longer to load than if not, assuming SNES core has the fast ROM loading functionality, which I think it does
But an update shouldn't change that unless that feature was disabled/broken
the situation is:
- Flashcarts inaccurately emulate some cartridge hardware
- Hackers develop against that inaccurate emulation
- They test on a flashcart and expect their hack works on original hardware because it worked on their flashcart (even though it wouldn't work if you put it on an actual cart!)
- More accurate emulators come along, and now and the hacks work on some emulators and not others
the correct point to break the cycle is step 3
yes but it's unrealistic to expect hack creators to go back and update every single hack that's now incompatible because of info they didn't know about
or step 2, fix it in high-accuracy software emulators as well
...maybe the incorrect SA-1 implementation from emulators/FXPAK should be considered as yet another expansion chip?
Perhaps, but it's really unfortunate if that holds back accurate SNES emulation going forward. How do we ensure that future hack developers are developing to an accurate target?
make the accurate behavior the default one in all emulators
exactly!
but still have a toggle for the old mode for compatibility's sake, is what i'm saying
Yeah, compatibility options are good where feasible, especially in software emulators
we already have an alternate core for just this feature, all i'm saying is that it would be convenient for that functionality to be merged into the main core
I dunno what you’re talking about but if you write a feature request using the GitHub issues it’ll probably be looked at
and I'm saying that'd be great as a disabled-by-default option if that's feasible! I don't know if it is for this specific change, maybe @elder pewter would know better
but sometimes, especially for hardware emulators, that's not feasible and you have to make a choice between accuracy and compatibility
i would be surprised if it wasn't able to be added to the main core with all of the other options it already has
was referencing this core which has different SA-1 behavior https://github.com/littlegamer87/SNESa1_MiSTer
i'm not gonna type all this out again ^
Wacky tabacky!
yeah i get it
the SNES core is very full, an entirely separate SA-1 implementation is definitely not feasible
but if it can be done as a small tweak somewhere inside the SA-1, then it's probably feasible
i was under the assumption that's all it would be; small tweaks. the repository states that these are the main changes
maybe srg could take a look at it and give a verdict as to whether or not it's feasible right now
hell this core already has a toggle for accurate or inaccurate behavior. not sure if anything else about it is different compared to the main
I am trying to catch up on this stuff, can someone summarise what the issue is?
Is it a situation where something new has been found out about SA1 hardware, that will be updated in the core, but in doing so a bunch of hacks will no longer work?
older emulators and the SD2SNES flash cart had inaccurate SA-1 implementation. many ROM hacks were (unknowingly) built around these inaccuracies, so they work on those emulators and flash carts without issue. when the SNES core for mister got SA-1 support though, the chip's behavior was better understood and was more accurate, but this accuracy actually breaks many of the older hacks in question
right now, there is a fork of the SNES core (SNESa1, linked above) that adds a toggle to switch between the new, accurate SA-1 behavior, or the old, inaccurate behavior. my opinion is that the toggle should be added to the main core and set to "accurate" by default, so it runs as it should for retail games, but those who want to play older hacks can set it to "inaccurate" to ensure older hacks run properly. nada suggests it isn't as simple as it may seem, though
I'm catching up but I believe the intention is to add it to the core if there is room after SRG does some SA-1 work #1096636309679919136 message
sorry if that was already linked to
thank you, i don't think it was. i hope srg can find the time between all of his projects to take a look at it 🙂
I hope he can work on it after he finishes the 3DO core! 💪

||I'm kidding! He should always work on what he wants, when he wants!||
If it can be done cheaply on a toggle then that is good. Hopefully someone goes and updates these old romhacks to work on the updated implementation as presumably this will be rolled out across flashcarts and emulators fairly quickly
Where has this knew knowledge a out the SA1 come from? I am surprised we are learning substantialy new things about the hardware in 2025
SRG started a deep dive on it
Oh, he really doesn't want to finish the 3DO core, does he?
I think he enjoys the suspense at this point
we are coming up on a year since he announced that
2023 i think, it was vitor viela or something who discovered a lot. not sure when book 2 snes dev manual was leaked on archive
Oh interesting. Was it part of the giga leak or whatever?
Nope, much earlier. I don’t remember exactly, but early 2000s
I get why 3D0 is moreso a project to chip away at when he needs a break. the 3D0 is just not all that popular
very few people could afford one back in the day
or justify spending so much on it
I'm really hoping there'll be some form of SFX3 on the core. if I understand correctly it basically only needs to be a SFX2 without speed limitations, so its only bottlenecked by the rest of the system.
Hello friends! I wanted to ask something—tell me, how does the Japanese version of Campus Challenge ’92 look? Compare the tournament version with my screenshot… I was only moving sprites around.
Is there anyone here who can help me make an accurate reconstruction of the Japanese version? The Pilotwings logo was hard for me to make… Programs I used: YY-CHR and Tile Layer Pro.
are you romhacking your dump to make it resemble the japanese version visually or something? I dont understand
We need to create the Picture in Picture video of yourself playing as well. Everyone start buying old analog broadcast equipment or you’re not even a real gamer
the cartridge never existed in Japan ? Hope someone it could be dumped
yeah looks like that is what they are doing
I have videos on my YouTube channel where I play Nintendo Campus Challenge ’92 and Powerfest ’94.
What I’m trying to do is make it look the same as the Japanese version used in that event.
The Japanese version does exist — it’s just difficult to find a real cartridge and to have the game menu dumped.
^^This would certainly be a great QOL improvement if all cores had this ability
Yeah it would be a nice toggle IMO
it sounds better than it actually is. there is a reason that OSD feature is there because this is risky
That’s going to wear down the SD card very fast because some games write to the sram area too frequently
Good luck to whomever tried that hack 
Ive always exited a core via the osd so this has literally never been a problem for me
I'll take that over the nonsense of corrupt SD cards any day
Yeah I think the simple QoL improvement of Sorg adding the "Saving..." text when opening up the OSD many years ago fixed my saving issues
fxpak has a background autosave, and we get fxpak users with dead SD cards all the time
it really sucks to troubleshoot, because it just manifests as weird inexplicable glitches
and no one ever thinks to suspect their sd card
What's the reason? MiSTer and some flashcarts are the only devices I know of that don't auto-write the save, just about everything else does.
There are SD cards that are meant for 24/7 video surveillance modern cards are a bit more robust than from many years ago, people run their entire OS install off of them for something like a Raspberry Pi. It would be fine if there was a delay to write or something for the games that just use it like additional RAM.
I don't think most o the people is going to use those types of SD cards? Anyways, having that feature as default and having no way to disable it is just plain bad.
I don't think they are even special cards. My point is cards are far far more robust now, many don't even use special cards, we don't need to worry that writing like 1K of data every few minutes is going to do anything.
Though I agree you should be able to disable it (You can in this, but you then need to manually save), but I feel in modern times letting it auto-write your save to the card really should be an option. ESPECIALLY for systems that used memory cards, it's really silly that I need to open the menu on the PSX core to save.
I'm pretty sure the eeprom chip used in PSX memcards at the time was less robust than modern sdcards
That's why I'd like to have it as a toggle perhaps with a warning
Yeah, I agree it should be a toggle with a warning, but I feel the option should be there
Let's see what Paul and Sorg think about that! 
No idea on what TempleOS is but 640x480 should be enough for everything 
Let's just say that's a hell of a rabbit hole
TLDR: Super religious person has schizophrenia and thinks God litearlly told him to code an OS that must be 640x480x16 (I think also must fit within 640K?) and that this OS will literally be some third temple that was mentioned in the bible... and it just gets more nuts from there
The OS was very impressive for one man coding it from scratch with such restructions, but like I said, HUGE rabbit hole
Oh the TempleOS story! I remember that and the guy is considered a genius by lots of people!
Was, he died 🙁
oh 🙁
He also became a lot less nice because of his mental disorder later in life sadly 🙁
Again.... 🐇 🕳️
twofold:
one, some games write to sram continuously, but flash memory unlike sram is limited in the amount of writes you can do to it. so continuous writes to your sd will definitely accelerate degradation. there is huge difference in the type of sd card you have
two, the de10 does not have the hardware for a shutdown routine, it just cuts off the power instantly, unlike the analogue pocket. since you have no control over when the mister is writing to sd anymore, that is a very big risk of data loss
They generally are special high-endurance cards. If you go to amazon or whatever and pick up a random cheap SD card, you get some bottom of the barrel crappy flash.
Most people just don't think about the quality of the flash in their SD card as being a thing they need
not everybody is a hardware engineer 😛 I get that
the safe shutdown thing is another problem fxpak users face -- you occasionally see SD card corruption unless you hold the reset button while you turn off the power, which signals the fxpak to sync
if you buy a high endurance SD card meant for continuous video recording, and you are sure to do a safe shutdown, then autosave is fine
agreed
but I'm a developer of the super metroid practice hack, and every few months or so we get a mysterious bug report that turns out to be an fxpak user who didn't do one of those things
the fxpak has custom hw tho, which should mean it is fixable. just add some capacity big enough to finish the write and that should take care of it
although the power that the snes can supply is probably limited, that can ruin things
you see those here too constantly. think it was only last weekend that I pointed someone here to test the sdcard. did not come back so looks like that was the issue
I've always been curious if those "endurance" micro SD cards would be better for something like that?
it can definitely handle more writes, but you will still have data loss if the power gets cut during a write
Yeah they are rated at higher amount of writes. Their MTBF rate is basically just exponentially better, but that still means it's a chance of wear and not a specific time when wear will happen after a certain amount of writes at a certain sector, it's just a lower chance for that sector over time.
Like I said, delay the writes to only happen like every second or so, that fixes the games that use the SRAM as actual ram trashing the card issue (Even though that should STILL not destroy the SD card. The person in the reddit post did some math and showed that running the SNES game with the largest possible save file nonstop having it keep saving would take 800 years to destroy a modern MicroSD card), and who is going to save their game then IMMEDIATELY turn off the mister before the second delay is even up?
Also if you run update_all often you are constnatly re-writing jotego's cores and now the timeline anyway
but you aren't running update_all constantly
It writes a LOT more data than a SNES save would though
Regardless, the delay I mentioned would prevent a lot of this
Assuming it even would effect a modern sd card without the delay like that core dev mentioned
There isn't any type of math that you can do to predict sd card failure. You can certainly set an expectation, but like any other wear and tear part, you cannot predict when it will fail - only that it will fail eventually
I mean, by that logic we should not use paging files on our PCs because they could destroy the SSD faster. That's going a bit extreme in worrying that auto-saving a few KB could kill the sd card
As long as we are not hammering it like it was an actual SRAM chip running RAM
it is sram, those games will likely be unplayable if you add a delay there because it also reads that data back
no wait thats not true, it is buffered in ram anyway
his math is off tho because it assumes that all capacity is available, but in mister it is permanently claimed by everybody for the roms, isos, cores and saves
do sdcards do wear leveling?
you're writing the same few cells over and over if they dont
@upbeat canopy in your opinion, why doesn't mister have larger adoption in the speedrunning community? It just feels like something like that would be easier than rgb and hdmi modding consoles so they can stream/play on a crt simultaneously
imo its because its not "real" hardware
thats gonna be the common answer lol
most speed runners are going to use analog signals on crts to boot, and they can capture video outside of the chain/digitally with say, a PVM that can pass thru video to a capture card
the capture is inconsequential, the important thing is that the crt gets the signal first, everything else downstream can lag as much as it wants
for Super Metroid, it's been cost + lack of SNI support
the original DE-10 board is comparable in price to a console + nice flashcart
the MiSTer Pi changes that, but without SNI support it still is a worse option than an fxpak for most people
That's fair. I think that will start to get better as more options become available. But yeah, the SNI bottleneck is real. Are you planning on PRing that for the SNES core? Or do you think it will forever be a fork?
I hope to upstream it
I want to send PRs for initial review by the end of the year, assuming I have enough time to work on it over the next few weeks
it just breaks my brain when goats does a metroid marathon stream and he is switching back between emulator and SNES.
is the SNI stuff just for the snes core? or would it be for the main binary?
from there it's a question of whether sorgelig and co. will accept it
https://github.com/MiSTer-devel/SNES_MiSTer/issues/422
Generally speaking this is beyond what MiSTer is targeted for. It has nothing to do with preservation. SNES core is already heavy core and i would rather to keep remaining room for main purpose - emulating original platform with high accuracy. So it's under a big question if this addon will be accepted.
i'm hoping i can get it into a state where it's acceptable to the project, but I have very little experience with VLSI design, and there's the question of whether they consider it worth the logic space
it requires:
- support in the core
- an extra user-mode background service on the linux side for handling communication with the host
- support in the main binary for launching the additional service
the core support is not terribly complicated, the only thing is it requires a lot of extra logic in the SDRAM controller to handle arbitration between the SNES and SNI (and that block is already pretty tightly constrained on timing)
the user mode service will be optional, I plan to distribute it separately from the mister project, and I've made sure it doesn't introduce performance concerns
I wouldn't really expect top runners to switch away from original console. They usually prefer the "authenticity" of it, and they've already invested the time and money into high quality capture so they don't really have a reason to switch
Rather I think MiSTer is a better alternative for mid-level runners who want something nicer than a PC emulator without having to spend $$$
that's totally fair. I just though it was weird that Eddie uses a Super NT (but, again, makes sense because the sd cart is doing the SNI work)
Yeah, mister has the potential to be a much better option than the Super NT
it's just that, right now, telling those people "you can't use an autosplitter or tracker or do multiworlds" is a tough sell
oh absolutely
I had in my brain that it was an accuracy thing - I didn't even consider the whole "hey, this whole hobby is uh...timing splits...and the mister cannot do that"
time... is of the essence
I'm hopeful that you can make it work in a way that sorg and srg approve of.
mister has a pretty big accuracy advantage over super NT because it can output at 60.09 FPS natively
with super NT you have to choose between losing like 10 seconds an hour, screen tearing, or variable input latency from double buffering
I know! I was really surprised that that was eddie's console of choice
did you manage to get this additional service allocated to the 2nd (idle) cpu core @upbeat canopy ?
i didn't do anything special for it yet
the main binary is already pinned to one core so the scheduler just naturally puts it on the other core
but it's easy enough to add a setaffinity call
it needs about 5% of a core under heavy load so i am very unworried about it
I meant a delay on writing the contents to sd, not the sram itself
What is SNI?
super nintendo interface
connect snes to a computer to read data in real time
some people use it to connect to apps to read memory states and keep track of splits and things when certain in game actions happen
like say, speed runners
Never heard of this before, do they solder wires to a snes or something? Or is it that bottom expansion port?
theres a few different ways
check the github repo for details
imo, its some super in the weeds nerd stuff, but the guys who use it, use it a lot. it's not really of interest for the average gamer who wants to just play games
alttp online is pretty interesting and does work on MiSTer to a degree thanks to NobodyNada #1096636309679919136 message
on FXPAK, it also supports Super Metroid, I think. haven't tried that on MiSTer yet.
I apologize if this is not allowed, but does anyone by any chance have this game? I’ve wanted to play it for years.
It was never released
i was just about to add https://www.indiegogo.com/en/projects/johnlester-4133569/justice-beaver-new-super-nintendo-game/comments looks like it never got out
Yep. Another scam. With compromise 
I read that as "with composite" and wondered why video signal was being brought into it
It has everything together, you can’t beat that!
Technically so does HDMI
It has everything together, BUT ANALOG
And you don’t have to license it. Checkmate hdmi
if only there was some kind of connector that could carry composite, RGB, and stereo audio all at once
Sharp scanlines as shaving machine.
With the age, I started to appreciate the real magic of composite.
Specially on Famicom ... I got an RGB famicom from aliexpress and it's great, but... Composite has magic.
OH LOL !
So which filter for the MiSTer is really close to that? Cause that looks REALLY good
Does anyone have any idea why yoshis island might be freezing for me as soon as I start the game? I would really appreciate the help if possible.
RAM issue? Or maybe corrupt save file
Are you on the stable core?
Yes I’m on the stable core
Try the unstable. There are some ram timing issues with certain configs and special chip games on the stable core.
You’re awesome! Thank you
@pseudo tangle ^^
Thank you very much for preserving these games! I just need someone to test the games on a MiSTer FPGA and upload videos, since I don’t have a MiSTer myself.
Does this mean it already works on the FXPAK PRO cartridge? Because that cartridge uses FPGA cores.
SRG320 implements those cartridge it will require update core and main mister
I don't know
I was talking with Mr. @midnight pumice and he told me that the games work very well. I hope he included the Powerfest ’94 1 million points version. I hope everyone enjoys these great competition games to play with friends or family—a Christmas gift from me and Mr. @midnight pumice.
In the main, the code identifies the header of each game and set how the core will handle those two cartridges
const char snes_cc92_header[] = {
0x00, 0x08, 0x22, 0x02, 0x1C, 0x00, 0x10, 0x00, 0x08, 0x65, 0x80, 0x84, 0x20, 0x00, 0x22, 0x25, 0x00, 0x83, 0x0C, 0x80, 0x10, 0x00, 0x00, 0xA0, 0x80, 0x01, 0x80, 0x80, 0x00, 0x01, 0x02, 0x2D};
const char snes_pf94_header[] = {0xC9, 0x80, 0x80, 0x44, 0x15, 0x00, 0x62, 0x09, 0x29, 0xA0, 0x52, 0x70, 0x50, 0x12, 0x05, 0x35, 0x31, 0x63, 0xC0, 0x22, 0x01, 0x80, 0xC2, 0x3A, 0x6C, 0xB0, 0xE8, 0x4A, 0x11, 0x20, 0xC0, 0xF8};
he sets also an option on the core OSD for the time
"P3oMP,Competition Cart time,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18;",
Wow, my respects to Mr. srg320. It’s great that he was able to add the different time settings just like on the original board. Remember that the original time is 6 minutes and 21 seconds.
why 21 seconds ? it is precise 🙂
Because that was the time Nintendo set on those cartridges.
The ROMs I shared on Internet Archive come preset with 6 minutes and 21 seconds built into the game.
I see thanks for the informations.
PowerFest 94 10K point is recognized but not the 1m points version (maybe link to the header if two roms have different headers)
@pseudo tangle Is it normal the headers of PowerFest 94 1m pts have this mention "PREHISTORIK MAN" ?
Honestly, I couldn’t tell you for sure—that was the original menu that Rick Bruns shared with me from his Powerfest ’94 1 Million Points cartridge. I only converted it into an .sfc file.
I will inform srg320 regarding the issue for this rom hopes it doesn't share the same headers of prehistoric man game 😅
rip Near
this post is from 2016
Could you upload YouTube videos testing the Campus Challenge ’92 and Powerfest ’94 games on MiSTer FPGA? I’d also like to see the option you added to set the time.
Sorry I don't have equipment for capturing...
@tulip vessel or @low basin ?
Thank man, I've sent you a DM.
I just edited the header in the 1 million points rom to match the 10k rom
@pseudo tangle what did you edit in the rom to set the time limit ? Unless the mister osd timer over writes it any way
The fx pak does not use the mister cores and will need to be implemented separately
SRG320 has updated the main mister to support 1 million pts rom.
I can take a look but not this week
Okay, I'm trying to get caught up on the competition cart stuff. So the carts are a collection of several roms? How exactly do they "connect" together so they can be played as they were originally intended at the competitions?
The roms that work are a single file, I've only tried the powerfest roms, but they just start immediately playing.
With the support of PowerFest 94 1m pts (Main Mister)
What does this file do exactly?
Oh, is that a test/unstable MiSTer main file?
yes it is. It is a PR not merged but soon as normally all is ok now
Nice, just didn't click in my head that it was a MiSTer main file.
sorry not to mention it on the message I will add it 🙂
So I was able to get all of them working except for Campus Challenge.
Felt like going back in time. I played in the PowerFest'94 and DKC challenges back in the day. Won a copy of Lawnmower Man for SNES in the PowerFest challege at the time...
im a dummy so apologies if this has been answered elsewhere but is there any more indepth info on what the new sound frequency option does
Simply put, some homebrews & demos seem to work better with some theoretical values on the SPC700.
With aren't realistic & never seen on the real chip, when measured.
Like with Bad Apple Demo or the title screen of Unholy Night.
i see. thank you for taking the time to answer!
was able to play PowerFest 94 with the "V2" mister main and the 20151218 unstable core. pretty cool to relive that memory 😄
Does anyone have the "link" or "search term" to find the competition carts? I have been searching google with no luck. Would love to play these games. Thanks guys!
Please do not circumvent the archive.org link ban.
I apologize—I didn’t know that wasn’t allowed.
No stress, thank you for acknowledging.
Sorry for getting you in trouble my man
what is the correct core/rbf file?
🤣
I found the games which should be left unnamed
Poor Link... 
So is the Campus Challenge that's listed as "original" the only one that works on the core? The other two I can't get working.
What core are you using? I cant get any of them to work lol
they worked for me with the latest unstable SNES in #unstable-nightlies and the MiSTer here #1096636309679919136 message
^ This. You need both the core and the updated MiSTer main binary. The core alone isn't enough.
I've gotten all of them to work except for the two alternative Campus Challenge files. Only the one labeled as original works. All the other challenge carts work fine on my end.
One of the alternatives opens to a different title screen than the one that does work, but doesn't go beyond the title screen. The last one just opens to a garbled mess of pixels.
that's how it was for me on the first alternative too, i assumed the point was showing the alternate screen only. i missed that last rom, looks like it was added later.
Yeah, there's one called second screen, and one called repro.
Original is the one that works.
I can get the other two working in an emulator on my computer. This is the one titled "Second Screen".
Doesn't work on MiSTer. Just garbled pixels.
second screen is the one that just showed that title screen for me, can't go any further
does it go further on emu?
Do you know if the official Analogue Super Nt core is still being developed? I noticed a bug today while using the German patch of Breath of Fire and I’m wondering whether it’s worth contacting Analogue support.
On the MiSTer, the patched ROM runs perfectly fine.
That really shows how good MISTer’s support is.
This is the one I can get to the title screen with on MiSTer, but it doesn't go beyond that.
That said, on the emulator I used, neither of them would go beyond the title screen either.
maybe missing headers on main mister I didn't this one on my side
With Analogue, no one knows. Sometimes years will pass and they will randomly release firmware updates, but they don’t communicate anything with their customers for the most part
I’ve just contacted support. I’m curious whether they’ll respond.
Here is the bug I discovered.
A photo of the Super Nt above, and a photo of the MISTer below.
It turned out that the bug might be related to the Flashcard FXPak Pro. I’ve now also tested the whole thing again on an original SNES.
Are you using the Ares emulator?
Looks like it's BSNES? I never play games on my computer (a Mac). I usually only use it to test patched roms before I move them over to my MiSTer to make sure they patched correctly.
Yes, I used the ARES emulator.
It is definitely caused by the FXPak Pro.
mister wins again?
As I mentioned, this is not caused by the official Super Nt core, but by the FXPak Pro flashcart. I tested the same ROM with the FXPak Pro on an original SNES and encountered the same bug there. Tomorrow, I will try installing the jailbreak on the Super Nt to see whether the bug also occurs there.
have you ruled out a bad SD card on the fxpak?
This is definitely not related to the SD card.
alright, as long as you've ruled it out (ideally by trying another SD card)
it's super common for fxpak users to encounter glitches like that caused by a dying SD card
On Internet Archive, I included the Higan and Ares emulators with their respective games. In the Higan emulator, the games are in folders, while in the Ares emulator, the games are in .sfc files. As an extra note, in the Retroarch emulator—both for PC and Android—you can also play these competition games; you just need the .sfc files and that’s it.
As an American I only use .snes files!!!!
Check out this video—I’m using it on my Samsung Galaxy Tab A9+, first with Winlator and then with the Ares emulator: https://youtu.be/ev4NaEqIbVc?si=XLXe88-4uZ9a3I-L
I’m telling Miyamoto!!!!
.snes files? Do those come from the SNES Classic Mini?
No I’m being an idiot, sorry
It will forever haunt you! 👻
The ghost of Robby past.
When does the act cease
its a shame that the fx pak/sd2snes cart does not get much love any more from its creator, if it was not for the community devs we still wouldnt have super fx/sa1/sdd1/sgb etc
The good point it is opensource for others to contribute
i thought it became closed-source after an older revision?
And it still dosen't have SPC7110, the website hasen't even been updated and still claims no SDD-1
yeah, this is a big one. FXpak is an extremely capable cart, but it's in a pretty sad state compared to the mister core imo
Latest MiSTer unstable build: https://github.com/MiSTer-unstable-nightlies/Main_MiSTer/releases/download/unstable-builds/MiSTer_unstable_20251221_076b2d
SNES: add Competition Cart support.
* SNES: add Competition Cart support.
* SNES: add PowerFest 94 1m cart recognition.```
Where do i put this file above?
Its not an .rbf file???
That's MiSTer main, rename to "MiSTer" and goes on the root of the SD card
omg, cool let me try that out
If using FTP you probably need to rename the current one then copy the new one over, rather than overwriting it
Then reboot
cool im on winscp right now just rename to "MiSTer"
Yeah, rename the current one to MiSTer.old or whatever and copy the new one over
Perfect it worked. Thank you so much. Im new but learning 🙂
NP 👍
i'm sure it's been asked already, but why are some core updates tied to mister main? you'd think they'd all be self-contained in their respective core
in this case, it's because the main mister binary has the detection logic for figuring out what type of cartridge a game uses and communicating that to the core
so new cartridge means that needs to be updated
main does the housekeeping, incl fpga programming and providing data like the rom and input
with some cores, main emulates a cd drive or a hdd
main also sends me usage data on what people are playing and runs my background bitcoin farming processes
You should put ads in mister main to get some extra revenue
See an ad for the prescription Tremfya before playing campus challenge 🤣
I had to look up Tremfya, I think I need some 😅
I'm starting to think the SNES core might be the most feature-rich hardware based emulator(for SNES)
NES core has got to be for NES right
Interesting. I knew main handled all the filesystem I/O but I didn't know it had a role in telling the core the cartridge type too
i assume the reason for that is just that that kind of autodetection is really easy to do in software and painful to do in hardware
Clearly the next route to go is to completely toss aside accuracy to implement gimmicks like HD Mode 7 then! 
I would love that as a separate core (SNES+?) of course it wouldnt be accurate, but perhaps more accurate than software based options.
On one hand, feels like we might be overdoing the alternate cores, on the other, we already have an overclocked n64 and psx core, as well as an alternate genesis core so...
i don't know enough about the internals of the core to say for sure, but a change like that sounds to me like a massive effort and very unrealistic to implement, even as an alternate core
the original PPU was very precisely designed around the target specs of the system, so things like the output resolution are fundamental to every part of the design, not just a variable you can tweak somewhere
all the memory busses, pipelines, and clocks are all designed around the exact amount of data that needs to move through the PPU each frame
in a software emulator, it's easy to write a custom renderer with whatever specs you want, since it really is just saying "ok now do 720 lines instead of 240"
but in a hardware emulator, that's essentially a complete redesign
and at the point you decide to do that, you're completely trading off everything that makes MiSTer more accurate than a software emulator. The whole point is to reimplement/recreate the hardware to get the exact same behavior; if you completely change the design like that then you've thrown that out the window and you're just another ordinary emulator when it comes to accuracy
Yeah we can only speculate unless we look under the hood. Some tweaks might be feasible.
But I'm a little reluctant to assume that everything completely is traded off. There are so many things that people were so certain about in this community, that they either had to backpedal or eat crow. So I'm comfortable with knowing that I don't know.
When I think of hardware mods, like HDMI mods for retro consoles, they have a separate board that taps into the original hardware. I imagine thats to some degree what goes on in the FPGA cores when it comes to features not in the original hardware, but I really don't know.
It's a lot easier to do something like that on a PC because you can just translate the console GPU instructions into instructions for the PC's GPU and just tell it "render these at 4k"
For FPGA you would have to basically create said more powerful GPU in the FPGA itself
How do you do an alternate core on MiSTer (like this one to play the comp carts) without screwing everything else up?
I think maybe there is a perception issue here. What are you worried about screwing up?
It’s hard to ask that question without coming off as condescending
I didn’t know if I can simply add the core to the lineup or if I have to do a major reconfiguration of my sd card
so the unstable stuff is kind of a misnomer. it's basically just beta. so if you run the beta mister main (which is what you linked), you likely won't notice anything AND you'll be able to play the champ roms. Easiest way to update mister main is to run this script (just download it by pressing the download button and put it in your Scripts folder - it will show up in the menu). That will update the MiSTer main file.
Then you can just download the rbf from the latest snes unstable, put that in /_Console, and run it
"unstable" just means "it might not work right"
like the SNES unstable currently has some graphical glitches in ALTTP because they core is mid-PPU-rewrite right now
To be fair, stable cores might not work right. 😏
I have never had any issues with any unstable cores, and if there were regressions the community is quick to find them and submit bug reports.
MiSTer has literally become the only place where I have zero concerns for installing beta software lol,
Commits · MiSTer-devel/SNES_MiSTer https://share.google/J4C5YTmBySaXr7ep3
better than 3do
Great update 👍
I wonder if this fixes the issue Nada was talking about
So did it fix a warning or is it a warning about a fix? 
The PPU rework is great stuff. first NES by Kitrinx, now SNES
So there’s no way to do it without taking my sd card out and adding it through my pc?
I’m in front of my mister with it on now atm
If you have an internet connection on your mister, you can use FTP.
Also these things are all just files. No need to reconfigure anything. Cores are just files, the menu list is enumerated at runtime, it's not preconfigured anywhere. If you add or remove files, it'll show up as soon as you refresh the menu page they're on.
The MiSTer main binary is a little more, and probably best to run that updater script that @woeful mortar linked to ensure it's updated properly.
Eh I don’t wanna screw something up by ftping (haven’t done it before lol)
So download the script in the link and put in in the scripts folder to run it once my sd card is back in the mister correct
Yes. That should update the mister main to the latest unstable. You may have to restart your mister after running it.
I feel really really dumb. Where’s the download button at
Nevermind I think I figured it out 😂😕
what are the benefits?
That’s what happened when I tried
Nevermind it worked 😁
I spoke to soon it skipped everything
Fuck
How do you find the unstable core to boot it once installed? I think that’s my problem
Latest cores are always in the pins.
It's wherver you put it when you downloaded it. If you put it in the _Consoles folder, you probably have multiple SNES cores, and it should have a different date
Oh, whoops, thought you were looking for it to download. To install it, yeah you need to put it in the _Consoles folder. And then just boot it from the usual place.
You can also create a new folder, like _Unstable or whatever you want to call it, alongside _Consoles, or even inside it, to keep them completely separate. Folders with an _ as the first character will be displayed in the menus
So wait…do I have to run the script to download it on my mister or ?
The latest SNES unstable is 2 things. The MiSTer main binary, which the updater script downloaded. You also need a SNES core, which you can obtain from the pins here
I have my pc on and my sd card in
So your SD card is in your PC right now?
I have my card completely set up outside of just getting the unstable stuff on it (been set up since I had it years ago lol)
Yep
Got the update unstable main sh in my scripts folder
You can find the latest SNES unstable core here #1096636309679919136 message. Download that, and place it in the _Consoles folder, or wherever you want as I described above.
@swift sierra do you see downloader.ini in the root?
Open it in notepad. add this line to the bottom [unstable_nightlies_folder] db_url = https://raw.githubusercontent.com/MiSTer-unstable-nightlies/Unstable_Folder_MiSTer/main/db_unstable_nightlies_folder.json
when you put the SD card back in your mister, run the mister unstable main script
and then run update all
it will download the latest unstable cores and put them into /_Unstable
and then you won't have to worry about it again
anytime you run update all, it will grab the latest unstable cores
and they will live in their own Unstable directory
yeah
Got it! Thanks 😁
Fyi for those booting it up for the first time, go into Hardware, and scroll down to Competition Cart Time to change the time (it defaults at the lowest starting at 3 minutes lol)
You can go all the way up to 18 minutes 😅
it does if you enjoy hitting baseballs for 15 minutes 😆
I think the default for the competition was 6 or 7 minutes?
savestate implemented? i cant check because i'm on a jouney
In any case, what a marvel of documentation we now have!
Seems to be this VHD file here?
A new .pdf shouldn't trigger an unstable build, if it has that is presumably a bug
No new SNES release in 5 months...
Depends on the auto-build backend. In the ideal world sure, have all kinds of ignores on .pdf .md .txt etc. so it wouldn't trigger a build. I'd imagine it's not that high up on the list of things to "fix" as it doesn't break anything other than wasting some compute resources to compile.
I think that is now the case, and this last unstable was triggered by an actual core update
Hi guys, I can't load any SNES or Super Famicom games, my screen just stays black. This romset has been working fne until I just tried it. I tried a fresh .rbf from github as well as the latest unstable nightly, nothing works. Any ideas? Thank you.
does the menu load? or when you load the core, it goes straight to black screen?
straight to black screen, i have to manually open the menu
and then booting any game is the same black screen, but you can open the menu?
yes, black screen, then I open the menu, select a game and just back to black screen. thank you
Try to load this
same thing, just back to a black screen
Does the core have been correctly downloaded ?
show me the root of games/snes
Html instead of a rbf?
Oh right on
I wonder if there is a rogue boot.rom
how can i show you games/snes?
You should have these two files in games/sense
any other files (that aren't roms) in games/snse should probably not be there
i have those, plus mister-boot.sfc
ok, that's ok. delete those two files and then run update_all.sh
ok, done, and still the same black screen 🙁
the one that doesn't work is HDMI (just SNES doesn't work) but I have a second MiSTer connected to a CRT using the same romset and SNES works on that one. which is weird
overwrite the snes RBF in /_Console with this one and try again (go back to the main menu first obv)
we this rbf did open the menu on boot, but opening a game is still a black screen
are the games on cifs? or on an sd card?
they're all on the same sd card as the system files, I have no extra storage hooked up
how about bring up the menu and reset settings
yeah, good call. it may be worth just blowing up the SNES file in /config
/media/fat/config -> next to nuke
if all fails -> get another sd card and do over. there's a slight possbility the card goes bad.
what do you mean by 'next to nuke'?
next file to nuke
do you mean delete SNES.CFG?
I think it does but we all know there is only one way to be sure
Deleting SNES.CFG seems to have worked! Thanks guys.
what is this
Powerfest 94
ty!
does the latest unstable have savestates?
I feel like I remember seeing that it does…
it does
and there is a boot room file that goes with it, right?
Yeah should be included as part update_all but if not I think I pinned it
Yeah it’s downloaded from update now
niiice
Maybe some day the release core will have save states
It's been on the expiremental core for, like a year now hasen't it?
save states? nah, it's only been on the unstable for 6 months or so
seems like years lol
The latest changes appear to be in a good spot, or at least I didn’t find any issues while testing a bunch of different games
So maybe release is soon
Feels like a long time to still be on unstable though
I'm the kind of person that uses only the stable version so the wait for SNES save states to be in update_all is brutal lol
I just created a seperate folder called _Expiremental where I put unstable and variant/special feature cores
... though the problem is cores that require an update to main
Never messed with manually changing that
Yeah the Saturn's need to use the latest unstable main turned me off from doing that. I know that the unstable main is usually extremely stable, but I don't want to take the chance. I definitely use the latest unstable cores if I'm testing though, but for playing for fun I only like to use the stable cores
SRG is very conservative about releases. Also, I believe he and Paul are in the middle of PPU rewrite
Not sure how to even do it. For the cores I just FTP into my mister and upload teh RBF into the expiremental folder
we likely won't see a stable SNES core until that work is done
Can I just ftp in and replace the main while the mister is running? Or does replacing the main or the menu require more than that
if you want run unstable main, just use this script. it makes updating it painless. https://github.com/misteraddons/mister_unstable_updater_scripts/blob/main/files/Scripts/update_unstable_main.sh
that will kill the mister process, download the newest mister main unstable, reboot your machine
Does that script only update the main or also the other cores?
Also, is it possible to add this script itself to the updater so that it keeps the unstable script(s?) up to date?
that script won't change and it only updates main, not the cores
if you want to update the unstable cores, there is something you can add to downloader.ini for that
I manually update the unstable cores in a seperate folder, I just asked because I noticed there was two scripts in that repo you linked, with one being apparently to update the cores. It wasn't clear if the main one updates both main and the cores, or just main
ahh yeah, the unstable core script is kind of obsolete, but the main one is not
Hmm, the description says "This Script will update to MiSTer unstable and the unstable cores"
Seems like a short script to update the main and other cores, it's not clear if it does both or just main
sorry, you're right, I linked the wrong script https://github.com/misteraddons/mister_unstable_updater_scripts/blob/main/files/Scripts/update_unstable_main.sh
that's the one you want - sorry
That's the one I was looking at actually
Hmm, looking through the code that I... semi-understand, seems likeit only updates main
oh then I did link the right one
I think the comment in that file is a mistake, I don’t see anything that downloads the mister cores in the body of the script, just the mister main
that one only updates main
I used it today 🙂
Nice, I'll have to add it.
Using an unstable main dosen't break older stable cores right?
(Or at least, it shoulden't unless it's a bug?)
right, it shouldn't
I am trying to remember a time where it did
I don't think there was one
Nice 👍
there was a time where unstable got pushed to stable too quickly and it broke saving in saturn
but I can't remember a time where games were broke by the unstable
Exactly lol
I can’t argue it, it’s smarter to do that. I’m just impatient lol
Personally I only vaguely am interested in save states but I have a friend who relies on them a lot and isn't interested in installing anything unstable
You can have both cores at once, that's what I do. I put all the unstable cores in their own folder
yeah I know he just doesn't wanna bother lol
save states is godsend. as adult i can't find that long duration of uninterrupted time to play games anymore.
Precisely. I feel the 3-4th gen in particular is an awkward era without save states since 99% of games don't have battery backed up saves
There's tons of games I want to play through, but don't want to be forced to do it all in one sitting
because I see a lot of Powerfest95 posts lately. Has somebody looked into Super Famicom Box support on Mister?
What's that?
That’s so cool
Hell no!!!!!!!! How dare you!!!!!!
yyyyyaaaaaaerrrgghhhhhh!!!!
you activated Robby's pirate mode
Y u mad bro
beep boop beep
One that clean would be a cool collectors piece.
Game capacity was…10 carts I think?
I have no idea, I literally found out about it when @left burrow mentioned it
Even though it appears like I’m a retro gaming encyclopedia who’s very handsome and strong, I actually learn new stuff every day here
also humble, I’m very humble
and cool
Yes
He’s also handsome and strong ||apparently||
yeah and is already emulated in the NoCash SNS emulator
It's a box! With Super Famicom games! It's the Super Famicom Box!
In the 90s, Nintendo took the guts of a Super Famicom, stuffed it in a black box with big cartridges, and sold it to Japanese hotels! Let's dive into one of the latest additions to my collection: the Super Famicom Box!
Additional links:
Super Famicom Box brochure: https://archiv...
and here is all about the tech; https://forums.nesdev.org/viewtopic.php?t=9597
what is nice about it is the Kirby Coin Intro and the Menu to choose the games from
I would love to see this working on Mister. I just got the Idea because the Powerfest Games work somewhat similar
I've had one for sale for 2 years and no buyers. $500 plus shipping or pick up in DC. I think I have 3 or 4 games with it.
Wow. Solid price assuming good shape and everything. Bet they’re north of $1000 now complete and clean
Pretty neat device, would love to see people study the hardware and rom
I bet there is unused/dummy data in there
You could always pay me £10 for 10 minutes of gameplay and partially experience the joy
Well if there is a dev interested I'd be happy to lend it to the cause.
that was done already by a few people.
If I understand the tech stuff correct, we should have enough information about it to make it work on Mister if there is still enough space for the core.
I assume that the menu was programmed into the main ROM, meaning it didn't have support for the operator being able to choose what games they wanted?
As far as I understand, there is the base rom with the menu/system stuff and the game roms. One game was always installed by default. The other games were optional. When a new game was installed, it was possible to choose it in the menu. There were only 4 games/boards. The last is very rare.
The menu lists the games though, and has graphics for them. That had to have come from the menu rom would it not? Unless the game carts were special ones that contained the menu graphic in them
If you know your list of games and won't change them later, you can include all in the menu rom only. But if you plan to add more games to the list later, you need to let the system rom at least know the name of the game and send the Game Logo Picture from the Cart to the Sys rom so that it can be displayed in the menu... in theory. Not sure what they did for the box.
Well yeah, that's what I was saying. If the list of games was set then they would just have the list programmed in the menu ROM and could use the commercial game roms/pcb design. If it's meant to let you change the games then the games would need to be modified to have a title graphic that would display in the menu added to them.
Has anyone spent a lot of time save scumming using save states on the unstable? I have been seeing some issues and I’m trying to determine if anyone else is
Basically, with the game will corrupt / crash when either I save a state or (more likely) when I load it. This happens on cifs and local (however, boot1.rom is always on cifs, so I wonder if that makes a difference)
This doesn’t happen every time, and I haven’t nailed down reproducible steps yet, but it happens at least once in a 2 hour super Metroid rando.
Things I still need to test
- does it happen on de10 based stacks (I’m playing on clone boards mostly)
- does it happen if boot1.rom (and the snes folder) is local
are there any consistent symptoms of the crash? e.g. if the crash usually happens right before a door transition, that narrows it down to likely being a crash on the SPC side
if it's a CPU-side crash and you're able to reproduce it in the SM practice hack, that'll give a stack trace
when you get a crash on loading a savestate, does reloading the savestate crash again? (i.e. is it a corrupt savestate or a one-time failure to load)
Oh it’s not the game the game I don’t think. That just happens to be what I’m playing at the time. It’s on save state side for sure, because the crash only happens when invoking the save state menu items.
Right, but knowing how the game crashes may shed light on what specifically is broken with savestates
The screen either completely corrupts when I try to save, or the saved state loads to a black screen
Hummmmm. I was getting it last night. I was running CPS2 core for a while and switch to SNES. Was either getting screen corrupt or game crashes. Only a cold reboot made the problem go away.
Sometimes you gotta clear it out like that with the MiSTer I've noticed
just in general
i always cold reboot when switching cores personally
For some reason I read that in a menacing voice 🤣
I’ve had this happen to me many times in the past. Switching cores without rebooting (even cold rebooting) often leads to issues like corrupted graphics or crashes.
Well that was quick! I reported the bug 3 days ago. 😂 Thanks! 👌🏻
fwiw, I haven't been able to trigger my weirdness on a de10 nano yet. I will keep poking at it
be sure to close your reports when you checked the fix 🙂
Is something considered fixed if it’s fixed in an unstable build and hasn’t been included in an official release yet?
Yeah
It’s fixed in the code but hasn’t been made public yet
indeed, request is fulfilled if the fix works
this organisation wrt stable releases is something the project can improve on, I think it currently doesnt always make sense, but I also recognise that it's not always easy
In Super Ninja Kun (Super Ninja Kid), the music of 5th, 6th and 7h boss stages dont load.
in both, public and unstable
@woeful mortar I have some trouble with the save state function on the last unstable version too, I can't load any savestate anymore.
Yeah, I saw just a moment ago about the issue ticket opened.
hmmm
And I don't see any file being created when I save.
That's weird, I would try rebooting because I can still save fine
and I haven't had the savestate bug on the new core
but there is definitely something weird in that department. not sure if it has to do with the PPU rewrite or not
I wish we had a sense as to milestone SRG is considering for another release. I haven't come up with steps to reproduce, and I don't have time to dig into the savestate stuff right now, but I also don't want to wait so long that it goes unreported
in new core, nothing is created in savestates folder qith me too
Very strange, only Super Ninja kun refuses to save a state.
Other games, like Super R-Type, Demon's Crest, 3 DKC are okay.
Super Ninja Kun dont load music of satge 5 boss
So, that game does something wrong.
yeah, SM is saving fine for me
I will add a comment about the broken save state system for that game, maybe it's an hint on what is going wrong.
I've seen several videos of Super Ninja Kun where the music is also missing during those specific boss fights.
Apparently, it's not only the core which is affected by that issue.
I tried all the save state TEST versions done so far, none worked for saving/loading with that game...
paul already figured it out
I thought he already implemented something for games with interrupts disabled, guess I was wrong
lets see how hard this game is and if I can make it to stage 5
great. the game also has an "enhanced music" (and sfx) option in the settings menu. if that is relevant this game is gonna be tricky
indeed no music. save state shared, I'd be surprised if this is not a game bug
i note 10 unused tracks in the SPC rip, so that would track
hm, i do see the music working in this video https://youtu.be/o1_9tJhq3tA?si=tMJES-a_A-BqQjbb&t=710
I have seen videos recorded from a real system and a SD2SNES flash card, which show that the music should be playing.
hmmm. this speedrun of the game seems to be done on hardware but music doesn't play in those moments https://www.speedrun.com/super_ninja_kun/runs/zp4e12rz
it's not marked as emulator and i see some telltale analog interference in the signal, it does look like hardware
yet there's no music playing in those boss fights
it's a game bug apparently?
according to that speedrun.com thread, you have to enable enhanced music in the menu or it won't play on those levels?
this probably needs one more test pass on the core with that setting enabled
I'm checking that video.
if you search the best games at the best price
visit the Facebook page "FROM JAPAN RETRO SALE"
Si vous cherchez les meilleurs jeux au meilleur prix
rendez vous sur la page FB de "FROM JAPAN RETRO SALE"
The video is done on a real system and a ream card, but the person have done an edit cut just before launching the main game...
Okay, I'm updating the issue ticket with the new information.
Thanks @sick crown
Let's do a second run, when someone is available to try it.
called it ffs
yeah your mention of a possible game bug is why i went looking for hardware capture that also lacked music
So, switching to Enhanced Music mode solves the problem?
ill leave that for the next person to try, I already called it 😛
it's not hard with unlimited continues
and you can increase your starting lifes from 3 to 7 in the settings
Is this one of the bosses to check? I lost count. I just turned on cheats to get here
Issue ticket updated.
Dumb question but does latest SNES unstable have the save states?
yep
yes but not for you, there’s code specifically to disable them for you
Oh good they're a crutch
Can someone test Super Family Tennis on real hardware ? After selection exibition on select screen, there is time to time glitches on transition screen with last unstable. I would like to confirm if it is normal or not
I opened two issue tickets about our discoveries.
Maybe, it's a rather recent regression, but I miss some time for checking previous releases.
I need to leave that task to someone else.
Scaler issue when switching resolutions, I asked for some help and where to address the issue on MiSTer structure.
i think sorelig is the only one working on sys
the scaler is in sys
if it is a digital screen, that can also be a factor. some screens take a fraction of a sec to change video modes
For my case, yes it's a digital screen, as I don't have the physical place on my desk even for a real PVM/BVM.
what is your vsync mode set to?
I have it set to 2 and I can even see that game switching mode before the title screen, after the namco logo
vsync_mode=1
same
your screens wont accept 2?
1 will match the refresh rate but has an extra frame buffer
that still image flickering is probably the buffered frame while it switches video mode/resolution
some other cores have an option to insert black frames during resolution switches so you dont see anything weird
oh that is clever
on vsync_adjust=2 I only see 1 "old" frame after a video mode switch
the elgato has way more trouble with these resolution changes than my screen has
the elgato basically resets the capture completely and you lose about 200ms
So, I recently learned there are carts you can buy for SNES that let you play GBA games in 4:3 (stretched, no border). I was wondering how possible it would be to get something like that working on MiSTer.
These carts have their own video out instead of using the SNES, which I expect would be a pretty big hurdle.
So are they even using the SNES at all?
Ok I looked into it, it’s called the Super Retro Advance Adapter and Retro-bit sold it. Other companies sold the same thing under different names.
It’s effectively just a GBA clone that gets powered from the SNES cart port and received button inputs from the controller. There’s no SNES software being run and no SNES hardware is used besides what I already covered.
So there’s no real point in replicating this because it has nothing to do with SNES hardware and it’s just bad clone hardware of a GBA, which we already have a cycle accurate core of.
the same company made a genesis "adapter" that functions almost identically. the SNES does nothing other than power the device and provide I/O
i messed around with one ages ago. it was not good lol
What would even be the goal?
I just wanna play gba games in 4:3 on mister, lol. A fringe case to be sure.
The Super Gameboy actually accessed the SNES hardware though right? I know it had a gameboy in it, but there were some enhanced games that were clearly running on the snes
could try setting up some custom modelines for the GBA core
Yep, that’s what I’m looking into now
Space Invaders is literally an SNES game on the cartridge when played via the Super Game Boy.
Yeah, that's the one I most think of
and odd thing is that SGB support for the GB version on the cart is excellent
Looks like Save States still not working on Mister Pi for Mega Man X2 on that unstable.
i believe games that use the CX4, OBC-1, S-DD1, or SPC7110 coprocessors don't support savestates
yeah, I believe that is correct
Yeah, save states are not yet supported on that chip.
No idea if they ever will be.
you could ask yourself if it is worth the logic if there are only 2 games for it
might have to wait for the de10 successor for that, leave some headroom for other dev stuff
I wish these two games were supported, but I know full well that it's quite difficult, and that's also one of the reasons why flash cards don't support them either.
The usual method used by paulb-nl for catching some interrupts can't be used there.
Sorry if I am asking a stupid question Do you use the turbo mode? In Super Bomberman for example there are very noticeable slowdowns if too much is happening on the screen so I use the turbo mode It’s almost unplayable at times without it
It is not worth the logic or the effort drastically speaking!!!!
At least FX and SA-1 are supported? Many big games used those
I would say WHAT games they are also matter. Nobody cared to implement those two missing SNES chips because they are just for some Japanese Mahjhong games, but something like CX4 is a bit of a different story
y'all, X2 and X3 have practice hacks with save states implemented.
there's probably a way to get just the save states...
might have to monkey with it, or maybe i'll ask the devs
So it's a hack that itself adds save states to the software?
yes
That's a bit different than the core doing it
X1, 2 and 3
well obviously lol
relevant if you want to use save states, though, no?
CX4 support for save states doesn't look like it is in the pipeline
which is why i mentioned it
they aren't ideal for casual play but they do have working save states, so presumably the save state support could be separated from the practice elements
I have been playing with the Chrono Trigger MSU-1 patch (https://www.romhacking.net/hacks/2546/) with resume support.
However, on the SNES core, dungeon music always restarts after a battle. So... Does the core support the MSU-1 "resume" feature?
Wonder if the version of that hack that will add in the PSX FMVs will ever be finished
Or if it's even possible
Resume support for the MSU-1 was apparently done here:
https://github.com/mrehkopf/sd2snes/pull/67
Is that supported on the SNES core? Do you guys happen to know?
I feel like I remember the music resuming in Final Fantasy 6? It's been a while though since I did that play through.
MSU-1 patched FF6?
@elder rose did the work on it as I recall, I don't know if he still hangs aroundhere
on the MSU-1 for MiSTer, I mean
Ah, let's see if he's around then. All games I tried seem to be lacking the resume feature on the core. I specifically applied an MSU-1 patch to Chrono Trigger that has resume, but doesn't seem to do it.
someone needs to add the FMVs from the PS1 version of mega man X3 into the MSU-1 hack. those are the only unique things left to the 32-bit versions
Huh, I remember X3 for PC... I think I even have it from years ago, don't recall it being on PSX
Wonder which version has the better quality FMVs
yeah, on playstation and saturn no less
@slow pumice the PC version has the highest-quality FMVs. the port included in the PS2 collection is about the same iirc?
Forgot there was a PS2 MMX collection, that had the MMX3 FMVs?
That collection was released on GameCube as well right? Would those be better quality?
Or the recent collections on modern systems and steam?
the newer X legacy collection only includes the SNES version of X3, while the X collection on PS2 and gamecube use the 32-bit version of X3. between the two consoles, i think FMV quality is about the same
Sounds close enough for having to re-encode it for SNES anyway
I could never really get into X3
I never really played X2 or X3, just X1 then went to X4
Really liked X4
... other than the script and voice acting that is
Yes. Using the Pixel Remaster soundtrack.
There are like, a dozen different options for MSU-1 soundtracks for FF6, I was using a patching program that lets you set what sets you want, it's deafault seems to mix different sets together
I tried several versions to test out and see which ones I prefer, sevearl options for the opera scene too, both with and without lyrics
I know something was implemented in the PSX core to prevent weird graphical problems during resolution changes when using the scaler. It sure would be nice if this fix could be ported to all other cores. I know SNES games don’t change resolutions as often as PSX games do, but it’s still very distracting whenever it happens.
The FF6 MSU-1 patch is called DancingMad (available in https://codeberg.org/Dizzy611/DancingMadFF6).
There's a single version of the patch.
From there, what version of the external OST (=PCM pack) is irrelevant with regards to features.
That said, resume doesn't work with FF6 using the DancingMad patch either, so it seems that, in fact, the core is missing the "resume" feature.
I will open an issue on github about this.
Just to be sure, and the exact same MSU-1 patched ROM has working "resume" on BSNES
So yeah, the core lacks the "resume" feature, making RPGs.. well.. less pleasing to play
I wonder how comes I am the first person to notice. Isn't nobody playing games on the MiSTer? 😄
We're supposed to play games on this thing?
No, no! How I dare? We have to worship it, all day long! 😄
A lot of people hate MSU-1
I only have one that I regularly play
MSU-1 is a tool.. hating it seems pretty stupid. There are many interesting patches. Most are pure detritus but some are incredible.
You're preaching to the choir
I didn't know the expression, but I was not preaching at all! 😄
What's that MSU-1 release you use?
Super Metroid Redux with the Symphony soundtrack
Ah! That sounds nice! I will take note of that one for when resume is implemented, as I tend to notice those things a lot and it takes me out of the game 🙂
I don't think I've noticed issues with it in particular
It's a missing feature, and since Super Metroid uses the resume feature quite a bit, I know I will notice it as I do with the FF games or CT
What would be the nature of the problem
In FF/CT, ending a battle causes the dungeon/environment music to restart, instead of continuing from the point where it was before the battle.
In SuperMetroid, it would manifest by restarting the music every time a jingle is played (which I believe is not what the original game did)
Ah I don't know if that happened then
Redux has less jingles too so harder for me to notice
is the resume command the only delta in msu1 v2?
I came up with something and compiling a build, see if it works
also jesus christ that patch is huge
Well, they're basically replacing all the synthesized sound tracks with essentially uncompressed audio
and a lot of it too, 1.4GB
I've opened a PR adding support for reading and writing SNES memory over UART for SNI: https://github.com/MiSTer-devel/SNES_MiSTer/pull/462
This PR adds a new UART mode in which the core listens for requests to read and write memory. This interface can then be used by a daemon on the Linux side to implement support for the SNI and QUsb...
@upbeat canopy can you close your issue for ALTTP? it is completed
@upbeat canopy This is very cool. I see you upstreamed the mister main stuff too ❤️. Is there still a separate daemon that needs to run on top of that. Or, if the mister main stuff gets merged, will everything be self contained to the mister?
I am really close at getting msu1 resume working
Yes, it's a separate daemon (at https://github.com/NobodyNada/snid). I think it's best to distribute it separately so that users who don't need SNI don't have to download it (it's 5-10MB), and so that the project doesn't have to take on responsibility for maintaining and updating it.
The Linux-side PRs just adds a menu option to the UART mode menu that looks for the snid binary on the SD card and launches it.
But I wouldn't be opposed to upstreaming the daemon if that was the preference of the project.
probably a little too niche, but I don't want to speak for everyone. Super thankful for your work here - really cool stuff
@onyx tulip can you give me the role to post in the test-builds channel? I got the resume command on msu-1 working
in the meantime, those who want to give it a spin, I uploaded it on github: https://github.com/MiSTer-devel/SNES_MiSTer/issues/461#issuecomment-3773566122
Hi there, The core MSU-1implementation lacks the "resume" feature, mainly used by RPGs. For example, in the original un-patched/vanilla Final Fantasy IV to VI or Chrono Trigger ROMs, when...
Seems to be working on my end. Tested it with FF6.
@elder pewter Thanks for your work on https://github.com/MiSTer-devel/SNES_MiSTer/issues/461, works perfectly and I am going to enjoy these RPGs with un-compressed OSTs a lot!!! 😄
you're welcome 😄 happy gaming!
Since it's tested, will you PR this to upstream?
pr is raised
Hey there, I've been away from all things mister for a while. Did the SNES core ever get save states?
I'd love to replay Secret of Evermore, but I wouldn't do it without save states nowadays
The unstable has them, but not the stable yet. They are in the middle of a rewrite of one of the components, so it may be a while before there is a new stable release
The unstable is working perfectly. I play it every day
what's being rewritten?
The PPU
interesting, i wasn't aware. i knew SA-1 was under review but didn't know about the main PPU
Yeah, I believe Paul found some issues with it (and fixed some stuff), and that activated SRG
at least, that's the story the PRs seem to tell
actually I take that back, SRG started it and Paul jumped in
the funnest thing about MiSTer for me has surprisingly been the little accuracy fixes
never would have thought those would be the things that got me the most excited when I first bought one
it's so cool
Nice! I'll give it a try. Thank you
I do still appear every now and then. Is this to do with the resume commit I saw go into the SNES core? Looks like Srg320 or Sorg sorted it out
I did see this and was a little surprised
Best teddy bear did it
No worries! And good to see ya here
I am lurking
reason I was surprised, is that the (un)official specification doesn't account for a resume
@finite saffron never added the control to the original test rom he created.
And I think he was right not to do so
rev2 with resume support is the current spec and should be added if it isn't currently included
ah, gotcha
we only did rev1 it seems
nice to see you @finite saffron , been a long time
hope you are well
Hmm so I wonder are there other features of rev2 that should be considered for addition?
I was under the impression that @elder pewter did the whole 2.0 upgrade, but maybe it's only resume? https://github.com/MiSTer-devel/SNES_MiSTer/commit/fefbccae4ea856ebab3c96e74ce0a91bf91145d7

