hckrnws

Check out as well:

- MiSTer FPGA project on GitHub: https://github.com/MiSTer-devel/Wiki_MiSTer/wiki (you can find there as well a link to MiSTer FPGA Discord with a section for developers)

- MiSTer FPGA in 2022: A Primer Guide to Retro Gaming's Hardware Emulator / MY LIFE IN GAMING: https://www.youtube.com/watch?v=rhT6YYRH1EI

- FPGA dev videos by Mister Retro Wolf: https://m.youtube.com/@misterretrowolf1464

- FPGA design books recommended by one of retro FPGA devs: https://archive.ph/Q30Xf

Almost all of my hobby side is in retro gaming, from having full SNES set to half complete NES and N64, a ton of old (micro) computers, everything from analogue, etc.. but I still haven't got into MiSTer. I thought it might be also a good DOSBox machine, along with other stuff it runs, but then it's kind of equivalent to what maybe 486 SX 33? If it were only a bit higher powered, let's say Pentium 120 or Pentium MMX 233.. are there a bit higher powered MiSTer machines? Is there even such a thing?

DOS gaming can be a bit tricky, it’s almost as if you need one for each generation of CPU/Graphics. Games like Ultima VII are too slow on a 386 but too fast on even a 75 MHz Pentium. Then you have the later DOS games near the end that use 3DFX/Glide.

I heard good things about 86Box to solve this issue:

https://youtu.be/qND3Eb0dI0E

Can't you just disable the Pentium cache and it runs like a slightly fast 486? I haven't done this but see retro YouTubers doing this often.

No, at least not yet. The limitation is not just the power of the underlying FPGA, but also the increasing complexity of the more powerful systems which need to be painstakingly recreated.

The MiSTer project specifically targets the DE10-Nano board and its Cyclone V FPGA SoC. There isn't a port or similar project available for some better board AFAIK.

> Also, I wouldn’t be doing my job if I didn’t mention all the red flags this has set off with everyone who’s ever manufactured products before; I’ve been creating and launching products for almost two decades now (at both much higher and lower quantities than this) and the information provided to us about this project over the last two months doesn’t line up with any manufacturing experience I’ve ever had.

What does this mean?

There was a time price of DE10-Nano was somewhat close to BOM cost, like one should do with your DEV demo boards to encourage experimentation and part use. But then people realized why design own product when we can just incorporate whole DEV board and here we are at >100% margin. Looks like this clone brings price back to those good old $100 times.

> Looks like this clone brings price back to those good old $100 times.

On the one hand, good improvement.

On the other hand, price and performance parity with a seven year old board containing a twelve year old chip? Oof.

Well, it’s mostly being used to simulate 30+ year old chips, so doesn’t need to be the latest and greatest

There are a few improvements that could still be useful:

i) FPGA on a newer process so it can clock higher. For faster 68k in the Amiga core and pentium 1 equivalent speeds in the x86 core.

ii) Faster ARM chip and better ARM-FPGA fabric, opening up hybrid emulation. Currently the ARM-FPGA layer is a real bottleneck and also the ARM core is a little slow.

iii) More spare IO to allow per-system custom io boards.

iv) Built in IO/memory board so they can fit in a nicely designed case rather than the current eyesore.

Some of those wishes may be addressed by the DE25 line, which is under investigation as a possible next-gen direction for MiSTer.

As discussed in https://www.youtube.com/watch?v=1Z0l9cLO0jA

>ii) Faster ARM chip and better ARM-FPGA fabric, opening up hybrid emulation. Currently the ARM-FPGA layer is a real bottleneck and also the ARM core is a little slow.

I'd hope for replacement with the open standard RISC-V instead, as used in many of the newer FPGA families.

The problem is that the FPGA is still priced similarly to when it was the latest and greatest.

I put off buying one when they were cheap because it was already old enough to assume it would be replaced. Then nothing replaced it and the prices doubled. Now that the prices are reasonable again, I look at it and think, something better will replace this soon.

One of the primary FPGA developers a while back said that it's unlikely that the MisterFPGA project will evolve to a much larger FPGA anytime soon because it's an immense gap between the current most complex cores and the systems that come after it.

Not only would it require a much larger FPGA, but writing the cores is becoming way more difficult.

The complexity comes from chip designers gluing IP together instead of designing everyday in house.

The good thing with stable hardware is that it is well understood, and for this task 'something better' isnt really required.

I also hesitated with the de-10 nano, but plunged in anyway a few years ago, with the MiSTer sdram expansion and some usb daughter board thingy. Blew my mind! and took me right back to real arcade machines. No lag, dropped frames, or other rubbish we just take for granted these days. All I was missing was a decent control deck, but I held off thinking I'd eventually build a cabinet, one day maybe!

I'm in the same boat, but this will do up to n64. What else could you want? Dreamcast? PS2? Not sure it's just power, and probably so much complexity vs PSX/N64 that it won't be done for years.

Yeah, there’s not much hope of anything anytime soon. Between Analogue and Polymega I’m covered for those generations.

All the FPGA based products went up like crazy in price during COVID.

Dreamcast ODEs like GDEMU clones went from ~45€ to over 100€, surpassing sometimes the price of one of the (almost never available) original. Or DCHDMIs over 200 or even 300€.

I know FPGAs got expensive during that time, but not THAT expensive.

I hope there are new generations of that kind of accessories done with low cost RPi Picos and the like...

Is there any indication which FPGA in the Cyclone series is being used? There are 40K cells up to 300K cells in this product line.

On the image from [1], I see 5CSXFC6C6U23I7N (110K). It seems to be rather expensive $300, 3 times the price of the whole MiSTer board.

[1] https://retroremake.co/products/mister-pi-retro-gaming-fpga-...

It's intended to be a drop in replacement for the Terasic DE10-Nano, and it uses the same Cyclone V FPGA.

The Digikey page indicates that the DE-10 has 110k LEs, so that must be the size.

I have very few complaints with the MiSTer, particularly from a technical perspective, however I have a large bone to pick with Sorgelig and other‘s on their forums, particularly in regards to cartridges. While the technical reasons for not using cartridges are valid (any implementation given I/O limitations would be equivalent to a Retrode, that is ripping a ROM + SAV only to rewrite the SAV and delete the rip), the childish anger I‘ve seen displayed towards cartridges is saddening and downright hypocritical. Have we really moved so far into a world of instant gratification that any minute not spent on entertainment is seen with vitriol and contempt? And yes, I know he‘s “seen these questions 100 times”, that’s also 100 times he chose to interact and waste everyone’s time being upset. If he’s not a social person, yet still responds to topics he doesn’t like, then that’s on him.

Personally, I can hardly wait for an open source cartridge-ready FPGA option to appear, and am particularly excited for Taki Udon’s projects. Till that point, the MiSTer is just a cute but incredibly expensive toy. Not yet a serious option for preservation.

What cartridges? Like for the Atari 2600 or Atari 800XL? I guess at some point you need to decide what type of cartridge to support because they have different pinouts and probably many other characteristics. Is there something like a generic cartridge interface?

> And yes, I know he‘s “seen these questions 100 times”, that’s also 100 times he chose to interact and waste everyone’s time being upset. If he’s not a social person, yet still responds to topics he doesn’t like, then that’s on him.

If he is not a social person, than I guess having to tolerate this kind of bullshit is much harder for him than for a more social person.

Sure, but then why consciously interact with it when you, especially as the admin, can just walk away?

Love seeing Retro RGB here. Bob does such amazing work.

It looks like a lot of work setting this up. Obviously it is a hobby thing as well, so I get that not every second needs to be 'productive', but it doesn't need to be this complicated.

I ordered an Anbernic RG35XX H a few months ago for about 60 euros and although it can't really handle playstation games, it does everything else. There are devices like it that can handle Playstation and N64 games though. I'd recommend this if you actually want to spend time playing.

Bonus: it is portable

The developer of this, TakiUdon, is working on portable devices as well. They have teased some lovely images on their twitter page https://x.com/TakiUdon_.

FPGA emulation is a niche within a niche and is targeted squarely at enthusiasts. I'm not surprised that you're expected to get your hands a little dirty to get it working.

I think it's mostly about the input lag. FPGAs will have a lot lower input lag so retro purists like it as it behaves closer to the real thing.

Well, great: "Up for sale today" (Sep/4). Unsurprisingly sold out now.

They were sold out in minutes, the next batch is going to be around 30 days after the current batch, so early October.

Looks like they’re already sold out of the first batch. It’s so frustrating to get into the FPGA scene, the Analogue Pocket sells out quickly as well. Guess I’ll stick to emulation.

FPGA based console emulation? There's been affordable (sub $200) and quite capable boards for ages: Arty T35 is $100-150 depending, and theres T100 models for more. although I'd check if a Zynq board is available.

And, vivado is free for this level.

The extreme learning curve of tooling AND designs that have few parallels to draw with arduino/uC code development. Along with the burdens ITAR creates- especially when installing the tooling (When's the last time you had to log into an installer??) are the grandest cliffs.

And much cheaper than anything else are the panologic thin clients. Still surprised more hasn't been done with those.

Roadbumps with the Panologic are connecting up a JTAG chain to the pins on the PCB and getting the VM containing the tools to run.

EDIT: Another roadbump is the need to have USB host IP in the FPGA to do I/O.

Given the multiple options that are now opening up - and there are more yet to come - the main market demand will likely be met within a handful of months at most. Handheld MiSTer clones may well be an Aliexpress commodity within a year; open source designs are already in progress.

None of this helps with the frustration of months of hype followed by a small run sell-out! but it does suggest that the wait won't be terribly long, from here on out.