FPGA Nintendo 64 gaming is here. 

Bob from RetroRGB should interview Robert once he is finished with the N64 core.

would‘ve loved to see more technical details of how the FPGA core was done

@@RandomUser2401There’s a little available here and there, but as you say consolidating it in one space. Plus an interview:) I love how he starts with a fully accurate software emulator (that he writes himself), and then moves to FPGA. His PlayStation core is super impressive to say the least.

@@flekkzoDidn't know his work but this dude's a fking legend if that's true.

does that mean we can make an ASIC and eventually a low cost n64 clone later

Would be interesting to see if Kaze Emanuar's work is too heavy to run for this setup.

​@magfal Kaze's work runs despite the limitations, so yes his Mario 64 hacks would work.

They had a really fast memory but random access and very small textures hindered it a lot it seems.

@@blahdelablah a lot of his work is focussed on keeping the memory bandwidth high by avoiding stuff that fits the hardware poorly. If an FPGA based system with too little memory bandwidth to 1-1 match the N64 is saved by most titles not being able to saturated the bus his work might be too heavy for such FPGAs

@@magfal You're missing the point I'm making. His work runs on real N64s. The MiSTer N64 core would (once it's feature complete and mostly bug free) have the same hardware capabilities and restrictions as a real N64. Therefore, if the code runs on one it'll run on the other.

I bet they can't do an FPGA of the cell processor before the year 2030

There are some hard clock limits on FPGAs. The Cell runs at over 3 GHz, and no current FPGA is that fast.

You know what they say, though - it'll be done. Not any time soon, but it will be done@@pikachuchujelly7628

Another fun fact is that the N64 supports memory-protection, at least on the main CPU, so you can have unpriviledged processes like with a modern OS

Surely a CPU is either a dual core or not? What made you say that, just out of interest?

@@aniluxneo9218I’m not OP and their answer will be more in depth but basically, is that the way a modern dual core works is by delegating the same task into multiple parts and splitting it up by core. The N64 had multiple chips that would do essentially the same same thing but with more dedicated and pre-determined tasks than a full multi-core processor from today would. Edit: Oh and because the co-processor didn’t have direct access to memory, it doesn’t work the same way either.

@@aniluxneo9218 They don't share memory-space, or a socket, and they have somewhat different instruction-sets

@@aniluxneo9218 the Nintendo DS, for instance, has 2 ARM CPUs both doing their own processes and hardware access.

True!

Librarians of the new age

No love for the pirates who dumped all of the console cart ROMs, and copied ISOs? ;)

@@videogameobsession so you are against pirating???

@@warrioroflight3489 It depends. If we're talking about someone who releases a rip of a Switch game, for example, before it even hits retail, then yes. If it's about making expensive games, old arcade boards, dumped, then no. I'm all for that. It's a fine line. Most often, people just want to be able to play games that companies refuse to, or are unable to release.

I believe intel subsidizes some of the cost of the DE-10 nano because it's meant as an educational tool. That's why I wonder where we go next because once we fully tap out the DE-10 nano the costs of this particular niche are going to explode.

@@devilmikey00It's not so much subsidized as it is that Intel's FPGA pricing varies wildly based on volume purchased. They will literally charge you 10x as much for a single unit as a volume customer. So Terasic is getting them with the volume discount and you're not getting that discount if you try to buy the FPGA itself on digikey. This has proven problematic for retro gaming product development where it's all happening at a small scale, and it's why you're seeing a mass exodus from Intel/Altera FPGAs to Efinix FPGAs. I understand that Efinix's development tools are pretty terrible compared to Intel's, but Efinix will match Intel high-volume pricing for low-volume orders and that's how you're getting Dreamcast on the MARS FPGA. Intel doesn't seem to have any interest in anything but massive volume customers for FPGAs, and they're probably going to completely lose that market (small volume customers) over the next few years.

@@guspaz They also charge wildly different amounts in different markets and Terasic is Taiwan and is able to get the full asian market pricing.

Can confirm to all who answered my original post that product availability and cost is extremely fluid depending on what tier customer you are. Low volumes? Get to the back of the queue. Nokia (back then anyway) - how high can we jump for you?

@@kristianTV1974 The price used to be much lower though. I bought my DE-10 nano for $135 in 2018 or 2019 (before the Covid), now the price has almost doubled. Still, from what I know, right now it's the cheapest FPGA for what you get for the price.

Idk if you've ever watched Kaze's videos but his newest video does a really good job of explaining how improperly managed polygons on the N64 can bring it to its knees. He renders one scene using 10000 tris and it it's over 30 fps but then he renders another that's 10 tris and the system slows to a crawl

@@clebbington As an avid Kaze enjoyer, which video was this? I'm interested.

I have no experience with FPGAs and had wondered about examples like the WWF game. I'm fond of Perfect Dark but that could slow to single numbers for frames per sec. Would exactly the same happen in a N64 FPGA core?

@@MrJakeTucker From my understanding, yes - FGPA is generally designed to be faithful to the original hardware, warts and all. There's some leeway with how you can control the levels of hardware emulation (for example with the N64 core, turning features of the VI pipeline on/off for intended visuals, line multiplication to fill a resolution correctly, and filtering on top of that for CRT effects etc), but there's usually not a lot of scope for implementing adjustments of clock speeds or simply brute forcing software to run faster. It's just mapping out how the chips work. In that respect, I think decompilation projects have a place in the scene for adding a larger amount of bells and whistles to a game, even if it takes more work per game and more copyrighted assets to do it.

@@MrJakeTucker Yup. If the core is doing it's job then everything right and wrong with the original hardware/software will rear it's ugly head. That said a lot of the newer Mister cores (including the PSX core done by the same guy doing the N64 core) have started to add a bunch of software emulation like options to smooth out the visuals, speed up the loading times and stuff like that. So, who knows, maybe there will be some stuff you can toggle on to fix certain inherent issues caused by the original hardware.

Who's "we all"? No commercial FPGA could handle the hardware of the N64 in the way it can consoles like the SNES. This is all just a glorified emulation, same as if you were emulating with a PC except maybe a bit more refined. If the hardware were correctly implemented you wouldn't be talking about implementing 3-point filtering because the hardware would be doing it by itself. Keep dreaming.

@@RaniaIsAwesome Really? Oh that’s disappointing to hear. What do you think of the potential for Mars Dreamcast FPGA?

@@RaniaIsAwesome"This is all just a glorified emulation" You think that's air you're breathing?

​@@RaniaIsAwesomeOr "tell me you don't know anything about FPGA without actually telling me".

@@thefunkdroid2777 💯 Dude has no clue about reconfigurable hardware. FPGAs do have limitations (for one, being able to “fit” the design) but calling it “glorified emulation” really shows you don’t know jack

Congrats on gathering 99% garbage. N64 has SM64 and that is it. Virtual Boy had more interesting games.

I was playing Saturn on my MiSTer this morning

That core is functional. It's not finished, but it exists.

Sega Saturn support is in a really good place at the moment. Some sound issues more so than anything else but a hell of a lot is playable there too. :)

Core's already out 🤙

What about Dreamcast we years away?

It's simulation, not emulation. :- )

@@HammyHavocyes this process would be considered that, but they called it the “emulation community” which is correct. Not a single person in the world would call this the simulation community.

@@BahhBahhBrownSheep I think it would be more accurately described as the "retro gaming community", or better yet, "retro gaming enthusiasts", as the "community", largely isn't. Most of the action is from a fraction of a percentage of people working on projects like this.

@@HammyHavoc Ah, semantics. It's what's left to discuss when all topics of substance have been exhausted.

It's _hardware_ emulation, as MVG said in the video himself. @@HammyHavoc

yeah

Retro gaming for me is just messing arou d with stuff like this, or emulators. I mean i love retro games too but now im more interested in getting them to work, than actually playing them

You will never be a woman

Yeah, it's sad how that trend likely won't continue for much longer, seeing as how the current and last gen's games are made.

Me too. I've put off watching this video, because "Analogue releases N64" may as well read "someone successfully implemented N64 in FPGA, but refuses to give it to anyone." So, AFAIC, whatever Analogue does is completely and utterly irrelevant. That so much of this video was actually about a MiSTer core is an unexpected, and quite welcome, surprise.

On top of that the only win Analogue will have for their N64 will be cartridge support & 4k output. If those two aren't important to you (they're not to me) then there's really no reason to buy Analogue's version. @@nickwallette6201

If they put in the work, they deserve the chance to use that work to their own profit, anything else is abusive.

​@@nickwallette6201I'm with you: open-source everything. BUT, as an open-source developer let me tell you that second to none of the people crying for a free solution are willed to give the devs some monetary compensation. Even worse in parts that are essential to the ecosystem, but only dependencies for other devs. No user will ever know about these guys and nobody will pay them a dime. The same thing applies to this implementation. If anybody gets a donation, it will be the MiSTer project.. not the dev of the core. Or even worse just some random shop in China selling that damn thing. Therefore it is nice that people like MVP give credit to the devs and enlighten some users.

@@absalomdraconis FWIW, I agree. I think Analogue would be much better off if they decided to reverse course and actually ... you know ... SELL the stuff they make. As is stands, you're either one of the first 50 people to get one, or you get to look at a picture of it on their sales page with no way to buy it. So, fine... you want this to be a theoretical product, then congrats on theoretically making an N64 clone. But don't expect me to care.

I designed a microcoded cpu on FPGA and honestly, once you get the building blocks together it’s really not that bad.

Chris taber is the reason analogue is so hated among retro console manufacturers.

I wish more talked about how garbage Analogue as a company. Like sure their stuff looks cool and works. Too bad scalpers are the only ones to get them all because of their shitty fake fomo system they have going on and their pre order system which is a scam.

lol wrong

@@MrAdam802 Do you really expect companies like Analogue to be around in a few decades? They sell expensive, incredibly niche products with a very limited production run.

@@MrAdam802 Cool take with all your evidence to back it up. Because we're all using BLEEM! to this day right?

wdym wrong

Back to Virtual Hydlide with you for now.

@@WhatAboutZoidberg yesh me lord.

And what a coincidence N64 FPGA became a thing soon after Ares was released.

Ares?

They haven’t, they developed their own hardware and software environment. Quit schizo posting because you irrationally hate a company.

Nope. Robert has already said that Analogue’s N64 FPGA developer has been helping him so you’ve got it backwards. UltraFP64 finished his N64 FPGA core years ago but he never released it since it wasn’t able to be used with MiSTer. There’s a good chance he took it to MARS FPGA or Analogue. Also, MarshallH works for Analogue to develop the Pocket and he may have brought his friends on board. You know the ones: together they decapped the N64 PIF and CIC to replicate them in FPGA and create stuff like UltraCIC and UltraPIF… which are used in products like the Everdrive and MarshallH’s own 64Drive.

@@jon4715 Ares is the latest open source multi-emulator that has an accurate (but still incomplete) N64 core. It's still currently the most accurate way to emulate the system though.

With the exception of the Duo, which is still plenty available for pre-order. Guess not too many people are interested in an FPGA TurboGrafx 16/PC Engine. 🤷‍♂️

@@DJ_Macphisto Which is both a god damn shame and not surprising. Only the Japanese would jump at the chance at getting a modernized PC Engine. It also doesn't help that literally the vast majority of the games are in Japanese only. There are fewer TurboGrafx16 games than there are N64 games worth playing. And that's saying something considering most people struggle to name 20 *_great_* games on the N64 lol (and by great, I mean holy balls amazing like what Super Mario 64 did for showing that 3D games can be epic when done right)

I can't imagine Analogue releasing their full working implementation before Robert finishes work on his Mister FPGA one, considering how fast and thorough he works. He is the same dude, who made Sony PS1 core for Mister and it was also extremely fast. Granted, PS1 is a more simple hardware design, but Robert Peip is a legend, I suspect the core will be fully complete or almost there by Christmas.

It was not bilinear but a Tri point thing that was not as good.

@@ivand5699 Tri point bilinear filtering vs normal four point bilinear filtering.

Breaks my brain to see an unfiltered 64 game. I love those blurry chunks of polygons

Fingers crossed for... [edit: Xbox next?] Just learned for GCN it's pointless - sorry lol: "...GameCube. Few people have access to one, especially the ones with 'Digital AV Out'/Component."

​​@@MachFiveFalconGameCube is probably the least needed fpga on earth lol. It is the single best emulated 3D console, full stop. "Authenticity" really isn't a justification, it's borderline OCD at that point.

@@-Monad- I forgot how accurate the Dolphin emulator is. I still have my old GameCube and haven't tried it. I guess original Xbox would be a better candidate for FPGA from what I've heard.

@@-Monad- Also, once you get into 3D people aren't looking for accuracy as much. Maybe some people want to capture that pure low res early 2000's look but I feel like what most people want out of 3D era emulators is upscaling to higher res, filtering and just generally making the games look cleaner. PCSX2, GCN and Wii all have incredible emulators that do that in spades and you don't even need a beastly PC to do it. Even on Mister the PS1/saturn cores comes with a whole host of stuff you can turn on to make games look less shit ;p Same with the N64 core. Wanting things to be look exactly as they were back in the day falls off hard once you get outside the 8/16-bit era's.

@@devilmikey00 i would argue that a large part of gamers actually exactly want that old low res early 2000s look. Have you followed streamers running older games like GC and N64? As soon as they dare to use high resolution or 60FPS mods there is an outcry in chat by people.And don't even dare to use improvements like filtering or anti aliasing. Even worse if you decide to use a high-res texture pack. People are really bitchy about those things and loooove to complain about it. Which is honestly annoying as then streamers are stuck having to play the "old style" that people want. At least those who don't just stream for fun but also have to keep their audience happy.

You people have like one joke

@@ireallydontknowifiamhonest you people?

@@JohnSmith-xq1pzyou know. The n-word: Nerds.

They are advertising 100% compatibility which is pretty bold if they only have an unfinished core. I assume they were working on it already with Kevin Horton?

Wait until the Chinese...

​@jimbotron70 That would be nice, but I have a feeling that if it was going to happen, it would already have done so.

@@Colin_Ames It's just a question of time, look at what they did with the OSSC clones.

@@jimbotron70 True. I will keep my fingers crossed.

Analogue N64 is a completely separate development than Robert's Mister FPGA implementation which MVG talks about here. They are two different platforms, right now only Mister's N64 is available.

Run operations in parallel and with the same timing as the original CPU/GPU and any other chips on the original hardware. Latency can be as low as the original hardware which generally isn't possible with software emulation due to having to process things sequentially, and also go through software abstraction layers like the DirectX API and an operating system as an additional layer of latency.

I bet Output, as rendering it higher could bring Problems with Sprites and Stuff that just isn't in that Resolution.

Presumably 240p at 11x Near Neighbor scaling but who tf knows.

They already said it would be for replicating different CRT masks but here’s hoping they also have a way to at least double the rendering resolution. I’m hoping for something like 3Dfx SLI where they might replicate two RDPs simultaneously with some kind of offset so that they render alternating scanlines that can be interleaved together for double the vertical resolution.

There is but they are really expensive and meant for development, something like the MISTer probably could be sold as a Pcie card but that' just an assumption.

Both the mister and the Mars projects are "hybrid", in the sense that they also have an ARM CPU, things like menu and processes that make sense make use of it. As for the "FPGA Card", there are many already, but the modern CPUs are already capable of doing low level emulation, offering practicaly the same thing as a card used for retro hardware would. However, PCs have a hardware and software pipeline GPU, OS etc to take care of, which add delays, so it's not as fast as a processing/display that's done all in FPGA.

Interestingly enough, these FPGAs do have a normal ARM processor inside them that is used in the MiSTer for the Linux backend that runs all the USB input devices, networking, storage, and probably anything else that isn't running on the FPGA side. However, it is not very powerful. You have a pretty cool idea though.

The next realistic breakthrough you have to wait for first is NAOMI/DreamCast. Hardware emulating The Cell, I’d imagine, would be crazy expensive. I think the later gens are benefiting more with software emulation anyways especially for QOL

ps3 has a very peculiar architecture, don't think we will see an fpga ps3 this decade

You'll be waiting a while. Dreamcast may be possible. I'm doubtful about the PS2 nevermind the PS3.

Its well beyond a DE-10 nano's (MiSter's) capabilities, but FPGAs available can already handle those consoles. Only limitation is how much effort it takes to implement a system that complex.

​@@carnivorebear6582And how much you're willing to spend on the FPGA!

OG Xbox is a Pentium 3 @ 733MHz and a GeForce 3. I think that may require some monster FPGA. For comparison, the MiSTer's AO486 core runs like a real 486 @ 25MHz with no 3D accelerator.

@kunka592 Oh yeah it'd take a monster FPGA, but they are out there - just incredibly expensive. I asked Kevtris a couple years ago in a MLiG livestream he was on what the limits were with current tech then, he said there existed FPGAs capable of emulating PS3 transistor wise, but they couldn't clock high enough to simulate it in real time.