Epic! 8-bitguy uses 1 weird trick to detroy rare prototypes!

To my point: I do not own an Amiga and I don't plan on getting one just for messing around, but I and many others do have calculators lying around. However I have been interested in tracker music for a long time and I haven't turned into a tranny yet so maybe I'll get back to exploring that too.
 
  • Informative
Reactions: Pee Cola
The problem (and honestly problem is maybe too strong a word) with these eight bit projects I could see is that the results at the end of the day just aren't very flashy and that usually leads to boredom and cessation of the whole thing. If that's not the case though and fun is had or a specific goal is in mind, then why not.

If anyone ever gets into Amiga hardware programming after reading this, just always go in it with the thought that this was a machine that was fundamentally designed around (analog) video, the likes of which we just plain don't use anymore. The real star of the production isn't the 68k, but the Agnus custom IC which contains a programmable coprocessor and a blitter. That's where a lot of the magic happens, without that the Amiga wouldn't have been what it was and it was all very clever and worked around some of the limitations of the time beautifully. That's what I meant in earlier posts when I said many of these projects get it wrong by being CPU-centric. All these older machines were about doing great things not because the fast CPU but despite the slow CPU by giving it a bunch of still very limited and expensive custom logic to do the lifting. Thing is, these days you can put something like the entire Amiga custom logic into a cheap FPGA with room to spare (and I cannot overstate this. I'm not up to what's up in FPGA land but if you make an economical design with an FPGA that's available for a good price you'd probably end up buying something you could fit several amigas into, these computers were PRIMITIVE) and it'll still be a lot faster than any Amiga ever was, so you don't really have these limitations to work around anymore. Technology progressed so how authentic can a new old computer ever be? Without limitations you don't need the cleverness. But enough about this derail.
 
  • Agree
  • Like
Reactions: Smaug's Smokey Hole, Never Scored, Pee Cola and 2 others
I learn because I like to fit things together, like Legos. I used to do a bit of x86 code golf and I wasn't ever great at it but it's interesting to think through a tiny program to try to find ways to save bytes. Sometimes I like to write in C just as an amusement to myself that I could. So it's ok to me if it's not flashy - I don't have any goals in mind.
 
  • Winner
  • Feels
Reactions: Pee Cola and ditto
AmpleApricots said:
I saw this union of Pi and ancient computer a few times now, even the Pi sometimes replacing the CPU and other capabilities of the system in question altogether. Kinda weird and frankenstein-ish imo, why bother with the old computer even at that point? But that's just like, my opinion man.
Click to expand...
I've seen an argument along the lines of "when does an old computer stop being an old computer?" where things such as FPGAs or Raspberry Pi expansions are added, but IMHO it depends on the old computer and how the Pi is used.

I agree that this line gets very blurry when considering something such as the PiStorm, an RPi-based Amiga accelerator that effectively replaces the 68k CPU. Granted, it doesn't touch any of the Amiga's custom chips AFAIK, but I understand why purists struggle with a mod of this nature.

When the Pi is used to complement existing hardware and open up opportunities to use a vintage machine without destroying its charm, that's a different story entirely. I'd argue that the various RPi addons for the BBC Micro fit into this category, as they're used either as a cheap way to do period correct things (e.g. the PiTubeDirect emulates all known BBC Micro second processors of the period, such as the Z80, thus allowing a BBC Micro to run CP/M) or as a way to make old hardware talk to new hardware.

The RGB2HDMI is a classic example of the latter point; originally designed for the BBC Micro, its low cost and simplicity has led to the RGB2HDMI being adapted for almost any vintage computer you can think of.

Another example of a Pi being used to complement existing hardware is the TIPI, which adds storage and networking to the TI99/4a.
 
Last edited:
AmpleApricots said:
The chinese will fake absolutely everything and they do not care. To be fair, some of their fakes are just as good as the original, but cheaper.
Click to expand...
Well yeah, where "cheaper" means they're cutting corners on materials, tolerances, and quality control. You might get a product that's just as good, but you might also get its reject cousin.
 
  • Like
Reactions: Pee Cola
If any of you are wondering what level of development hell the X16 is in. It's here:
www.commanderx16.com

Change of product direction, good and bad news!

I wanted to address the recent departure of Perifractic and talk about some of the change in direction. This could be a long one, so settle in. The absolute first thing I want to clear up just in case anyone is thinking this, Christian (Perifractic) and I are not mad at each other, we didn't have...
www.commanderx16.com www.commanderx16.com

Posted this back in august but never made a YT video on it for some reason.
 
  • Informative
Reactions: ditto, Pee Cola, Smaug's Smokey Hole and 2 others
speaking of David, he recently made a new video about the comparison of the 1983 Speak & Spell to its 2019 rerelease:
 
  • Like
Reactions: Pee Cola
Coolio55 said:
If any of you are wondering what level of development hell the X16 is in. It's here:
www.commanderx16.com

Change of product direction, good and bad news!

I wanted to address the recent departure of Perifractic and talk about some of the change in direction. This could be a long one, so settle in. The absolute first thing I want to clear up just in case anyone is thinking this, Christian (Perifractic) and I are not mad at each other, we didn't have...
www.commanderx16.com www.commanderx16.com

Posted this back in august but never made a YT video on it for some reason.
Click to expand...
I think he did mention it in his Attack of the Robots video somewhere.

As someone who writes code and knows next to nothing about EE, working with hardware is incredibly frustrating and takes a lot of knowhow and experience. I can't imagine how difficult it is to ship a new product that is anything beyond trivially letting PCBWay handle everything. I know people who have done the PCBWay route for personal hardware projects and the result is very cool, but this project was way too ambitious from the start.

And what value it provides over widely available Arduinos, RasPis, even TI calculators is not great. All those have established communities and so toolchains, emulators, help sites, etc.
 
  • Like
  • Agree
Reactions: Karma Khan and Pee Cola
Coolio55 said:
If any of you are wondering what level of development hell the X16 is in. It's here:
www.commanderx16.com

Change of product direction, good and bad news!

I wanted to address the recent departure of Perifractic and talk about some of the change in direction. This could be a long one, so settle in. The absolute first thing I want to clear up just in case anyone is thinking this, Christian (Perifractic) and I are not mad at each other, we didn't have...
www.commanderx16.com www.commanderx16.com

Posted this back in august but never made a YT video on it for some reason.
Click to expand...
What's going on with Perifractic? Anyone knows what the "personal issues" are?
 
  • Thunk-Provoking
Reactions: ditto and Pee Cola
You know, if you have to preface your statement with "I ain't mad at" for a damn product update, I kinda question your integrity. Like, you're just barely shitting on them at this point, just not at the petty level you'd see from other people intruding the hobby space.
 
  • Agree
  • Thunk-Provoking
Reactions: ditto, Pee Cola and Pissmaster
Kosher Salt said:
Well yeah, where "cheaper" means they're cutting corners on materials, tolerances, and quality control. You might get a product that's just as good, but you might also get its reject cousin.
Click to expand...
Or sometimes, you just get a superior knock-off that improves on the original product in sensible and good ways. Until everyone buys one over the original product (as sort of an insider thing) and business of selling that knock-off goes well until some chinese smells he can make 10 cents more per piece and replaces e.g. the die of a knock-off MCU with spoiled mustard destroying the budding confidence in the knock-off and making sure nobody ever buys one again, while the chinese in question made a negligible additional profit and actually lost money in the long run as opposed to if he would've just played it straight with the western market. That's very oversimplifying it but that's pretty much China in a nutshell but you can write books on this and this is neither the place nor the time.

Pee Cola said:
I agree that this line gets very blurry when considering something such as the PiStorm, an RPi-based Amiga accelerator that effectively replaces the 68k CPU. Granted, it doesn't touch any of the Amiga's custom chips AFAIK, but I understand why purists struggle with a mod of this nature.
Click to expand...
The thing with these accelerators (I'm talking original/enhanced chipset, although the same applies for AGA just with different numbers) that everybody loves so much is that they're often kinda pointless except for benchmarks. You can speed the CPU up a lot in an Amiga and even give it local RAM it can talk to directly exclusively, but at one point it's gonna have to talk to the chipset and it's RAM and that runs at ~7 Mhz if you don't move it into a faster FPGA. Since everything is derived from a main clock you can't really overclock it either if you don't want to end up with a machine that simply does everything wrongly. If you run your CPU in unfortunate ways at odd clocks bad timing and missed "windows" in which you talk to the chipset can actually make the computer as a whole behave slower in chipset-dependant operations as if you just had a slower CPU and beyond a certain CPU speed that wonderful Agnus with it's Copper and Blitter turns from a big help for the CPU to a lead anchor doing things painfully slowly the CPU can do much faster just by itself. Programs written with utilizing them fully in mind (so almost every program ever) won't even run faster then since your CPU is just twiddling it's thumbs (and screwing up hand optimized timings in the code, sometimes leading to odd bugs) until they do their thing. If you're unlucky the programs will run *even slower* than with the original CPU with your Accelerator if asynchronous timing between CPU-land and Chipset-land don't fit well and we miss windows to talk or even have to make one half wait for the other, with both doing nothing.

Then at some point and with the the right software it becomes a game of "how fast can I stuff that chipset-accessible ChipRAM with the contents of my CPU-local FastRAM" where I do everything in CPU-land and basically just give that slow chipset the results and things get frankly plain silly. And again, remember that there's a limit because that Chipset has one set speed derived from a clock that rules everything from floppy disk R/W speed, to VSync, to audio playback frequency. It doesn't matter if your CPU is 50 or 500 Mhz. The sweet spot for a dream OCS/ECS Amiga CPU is 14 Mhz (double the chipset speed, so we get a CPU-side bonus cycle over the chipset to think things over but are still easily synchronous without excessive cleverness) and some local FastRAM for the CPU to access exclusively (and we need that CPU local fast RAM else we'd just waste that extra cycle to wait for chipset access, speeding up nothing as the 68k has no internal cache) That's a nice and very noticeable speed-up without going overbroad in complexity and the Amiga would've been a so much better machine if it would've just been designed like that originally, but well RAM and especially faster-rated RAM used to be easily the most expensive thing about a computer and that design wouldn't have made much sense without extra RAM. (although giving an un-accelerated Amiga extra CPU-only RAM is worth it, so the blitter can do chipset RAM-bound stuff while the CPU does CPU RAM-bound stuff, otherwise they'd have to wait their turns)

Everything above that is kinda diminishing returns and "why would you even try that on an Amiga" territory, at least as far as the original 16-bitters go. AGA has more bandwidth but you're just gonna run into the same bottlenecks at different numbers.
Of course you could also redesign the chipset in an FPGA to simply be faster at some points with faster RAM and everything but man.. at that point, all that work for what? For high color desktop backgrounds? To run Elite II kinda faster? To make the floodfill in Deluxe Paint III faster? Kind of a waste of time if you ask me. People won't suddenly start developing for the Amiga again, especially not if they first have to buy your accelerator which costs as much as a new mid-range Office PC.

I mean I'm not gonna lie and of course you'll profit somewhat linearly off a faster CPU in CPU-only loads which is all of programs at least some of the time (and I have used and own everything up to including 68060 CPUs in Amigas, got them at times where nobody cared about them) but it's not as simple as it maybe seems and I can guarantee you most people buying these accelerators don't even know what I just told you because they absolutely don't care about Amiga internals and just want to consume product and have the longer benchmark bar to prove that they successfully consumed product. I just googled the PiStorm out of curiosity and they recognized that bottleneck problem too apparently because they put the graphics right on the Pi itself via RTG, (ReTargetable Graphics) bypassing the Chipset in that central area the Amiga is all about completely. (which software has to be explicitly written for or at least be very OS compliant, a ton of stuff like games directly bitbang the chipset and don't and won't ever work this way) Why even stick the Pi into an old Amiga at that point and not just run an emulator on the Pi itself? The Amiga basically isn't really utilized anymore anyways and it's easier.

You could already hand-build such a simple 14 Mhz accelerator in the 90s. Nowadays it shouldn't cost more than <$100 (although the right +5V RAM might be hard to get) and is a good "immediate EE hobbyist" project.

Good example why these new computer kits are such a different beast from how these old computers were designed and what you had to think about.

Pee Cola said:
The RGB2HDMI is a classic example of the latter point; originally designed for the BBC Micro, its low cost and simplicity has led to the RGB2HDMI being adapted for almost any vintage computer you can think of.
Click to expand...
This is mostly how I like to keep it. Have external modern hardware to talk with the mostly untouched old computer. Like I said, I use a small ARM SBC (that would've been an absolute supercomputer superpowers would've killed each other for if it was the 80s) as a *nixbox and modem emulator via serial. I use modern FPGA hardware in the form of an OSSC to get the computer connected to my modern screens. I'm not opposed to that. I also make an exception for drives, mostly because original drives of that era are either dead or would be a PITA to work with, even though often the MCU in them is usually a ton faster than the computer they're connected to. I mostly just wouldn't do things like the PiStorm for the spoilered reasons, the old computer tends to just end up being an useless appendix you could just leave out of the equation and even have a better experience for it.
 
  • Informative
  • Agree
Reactions: Cowboy Kim, ditto, Basic Block and 3 others
Coolio55 said:
If any of you are wondering what level of development hell the X16 is in. It's here:
www.commanderx16.com

Change of product direction, good and bad news!

I wanted to address the recent departure of Perifractic and talk about some of the change in direction. This could be a long one, so settle in. The absolute first thing I want to clear up just in case anyone is thinking this, Christian (Perifractic) and I are not mad at each other, we didn't have...
www.commanderx16.com www.commanderx16.com
Click to expand...
The more I read, the more confused I get about what the Commander X16 is supposed to achieve. Especially the Phase 3 and X8 derivatives, which sound like FPGA-based SBCs.

Right now, my opinion is that the most obvious way for David to save the Commander project is to try and get the X8 down to a price that's competitive with the Colour Maximite 2 Gen 2, only in a wedge style case similar to the Pi 400. Granted, it won't allow for the best keyboard, but it'll still be a hell of a lot better than many of the keyboards in 1980s computers.

If priced at say $100-$150 and with David's name attached to it, the X8 might end up selling 10,000-20,000 units, which would no doubt lead to a vibrant and active development community. If the ZX Spectrum Next has managed to build a strong developer base with just 3,000 machines out there (expanding to around 9,000 once the Issue 2 is delivered), then there's no reason why the X8 couldn't replicate this success. Heck, I'd probably buy an X8 to go with my CMM2, Pi 400 and (yet to be delivered) Next.

As for the X16... pitch that as a more expensive niche product, possibly in kit form. They might shift a few hundred units. I'm still not convinced it'll be the best educational 8-bit computer out there, nor the most useful... but again, The 8-Bit Guy has a brand name and some people may choose the brand name over something that doesn't (such as a CERBERUS 2080 or RC2014).

I really think that David and his crew both underestimated the work required to build a new microcomputer from scratch, and overestimated demand for the same.
 
Last edited:
  • Thunk-Provoking
  • Agree
Reactions: ditto and Smaug's Smokey Hole
Without a business liaison in China (yes, actually somebody being physically there. This is important) looking after the various Chinese manufacturers and suppliers and checking that everything's on the level they would've been ripped off anyways. Also with some countries, trying to import a fully functioning electronic device for the non-professional end user vs. a bunch of parts is a very different thing customs wise, are they even aware of this?

He should just get PCBs from china and sell kits, otherwise he's going to get fucked especially if 4-digit sums are already painful.

E: Also already talking about "possibly ASICs". Jeez.
 
Last edited:
  • Like
  • Feels
Reactions: ditto, Hellspawn and Smaug's Smokey Hole
Mr. Duck said:
None, David is a Commodore autist and basically wanted a C64 on steroids.
Click to expand...
Sounds like what he actually wanted is a MEGA65, only cheaper and with discrete chips rather than FPGA. The MEGA65 has only just gone to pre-order stage after being in development for 7 years, and it's around $750.
 
  • Like
Reactions: Hellspawn and awoo
You must log in or register to reply here.
Back