Accelerating the STEALTH Apple IIGS! 

need to look into one for one of my IIGS's. Have fun

Where can we buy these?

14MB?! What a madlad. If you have any videos about it I'll check it out ;)

that’s so cool Niek

Schematics or it didn't happen 😁

Yes! This needs to be a memo to all retro computer channel dealies. Feels much more 'hands on' and authentic - gives me a feel for the machine.

Agreed. Hard to work on the machine too though.

I'm amazed he doesn't have one as some one that messes with these old computers. I had one when I used to make genesis repro carts lol

Jobs had nothing to do with it, blame WDC. Getting samples of the 65816 running beyond 3-4Mhz in 1985-86 during IIgs development was a problem. Even in 1989 when the TranswarpGS came out it was an issue getting 8Mhz samples running. The early cards were clocked around 7ish Mhz and had GALs to slow down the CPU during certain instructions due to bugs. It really wasn't until the modern Sanyo produced CMOS 65c816-14 that "good" CPUs became readily available. Those units would clock at 20Mhz without issue in many cases. The CMD SuperCPU for the C64 used these. Of course this was long after the IIgs was current.

Classic Jobs. At least his random quotes are fading out. In favour of Elon Musk. Yuck.

@@DavideNastri at least Musk didn't con his friends. I hope 😁

@@akkudakkupl no, just his business partners!

Do it! Good luck!

Oh man, the would be so disappointing

For most of us, snapping a capacitor is a learning experience on sourcing & soldering & the occasional botch repair. I'm sorry you didn't grow up with the cool kids at the radioshack

ReactiveMicro can reproduce the cable

Nice score

Zip Technologies went the cheap to produce route of cramming the slowdown logic into a ASIC. The TWGS went with an early CPLD and discrete logic, the bonus of which is that it can be more readily upgraded, overclocked, and repaired.

@@NJRoadfan Hard to tell if it was the 'cheap' way. It was a pretty good accelerator and could easily archive 12,5 MHz at the time. With the later 14 MHz WDC 65c816 chips, the ZIPGSX can go up to 16 MHz in most scenarios (it needs selected 12/15ns SRAM for that to work). The ASIC is actually running at a constant 20 MHz with overhead 16Mhz is most likely at the upper end of technical possibilities. The 65816 CPUs at the time were buggy and reached 10 MHz at max (overclocking to 12 MHz was possible with selected chips). I have tested my defective cards and came to the conclusion that it is most likely the ROM in the ASIC that went bad. It may have a limited number of read-cycles before it dies. Sometimes heating the card to 150°C for 2 hours in the oven brought a card back for a few months... but eventually all cards died. I am the first to say, that the Transwarp GS is the better card despite the older technology and less efficient design. The TWGS is just a lot more repairable than the ZIPGSX. I have 3 TWGS and all work normally, one is even stable at 18 MHz (about the same speed in applications than a 15-16 KHz ZIPGSX). I would like to see a revamped TWGS with a modern FPGA at 20 MHz. It should be possible since the code for all the programmable chips on the card is known. The ZIP Tech guys were very good at designing cache circuits and were very good at large scale integration. The 6502 ZIP-Chip was a marvel at the time. It used an integrated ASIC, cache and CPU on a ceramic (later normal PCB) carrier in a "chiplet" style configuration. Very advanced at the time. I suspect something similar for the ZIPGSX ASIC and it may contain multiple chips/chip-dice. Maybe the 40pin 'debugging'-connector was used for programming the internal OTP-ROM? Here are some internal images of the ZIP-CHIP 4000 and 8000... as well as a lot of 6 defective ZIPGSX before repair. drive.google.com/drive/folders/0B5SAWSGa49rLamJSZTloVkphQjA?resourcekey=0-NPOtW0st5FVpAidNBKirPQ&usp=sharing

@@john_ace Some awesome images of the ZIP-CHIP internals - thank you!

Yeah, you show him who is boss! 😉

The IIgs (accelerated or not) might not be able to keep up. The video RAM on the IIgs runs at 1MHz, and Wolf 3D on the IIgs works by rendering the game first to fast memory (which runs at the speed of the CPU) then using the IIgs' RAM "shadowing" technology to copy the fast RAM to slow RAM with a technique called "PEI slamming" - basically the program has to copy the graphics data onto itself to trigger the "shadowing" feature. To do this as fast as possible, the program relocates the stack to coincide with the location of the graphics data so it can use a 6-cycle instruction (at 1MHz, because it's shadowing to slow RAM) to copy each 2 bytes of display data... So to "PEI slam" the whole display would take 96ms (frame buffer is 2 bytes per 4 pixels, so 1.5ms per pixel at 320 x 200), meaning that if the game were doing *nothing* other than updating the display from the double buffer, it could only get 10 frames per second running at full screen. Regardless of what kind of accelerator is in a IIgs, that's the speed limit for updating the display. As shown in the video the game is using around 30% of the display, so it should be able to update from double buffer in around 29ms, increasing the theoretical upper limit for the game's frame rate to around 33 fps. (EDIT): I'm not entirely sure I'm correct with my analysis of the timings here. the PEI instruction takes 6 cycles to read 2 bytes of memory and write it, in this case, back to the same location. Since it's a write to shadowed RAM the CPU does need to slow down to accommodate the speed of the Mega-II core. The timings I calculated supposed that the CPU would be clocked down to 1MHz for the whole instruction, but I've seen info that seems to indicate that it would've just had to wait long enough for the Mega-II to sync the update to slow RAM, so maybe PEI-slamming is faster than I thought (Somewhere between 2-3 times faster than what I calculated, by my guess, since only the "write" portion of the PEI instruction needs to run slower)

It all has to do with the tight coding talent of John Brooks.

I am curious about this myself. I suspect that part of it is that we perceive the speed issue more in Super Mario Bros (a kind of simple, but fairly fast-moving game) than in Rastan (which should have buttery-smooth scrolling, but the game itself is slow enough that a bit of chop is maybe less noticeable)

I think it's rarer. The TWGS was just a more popular product at the time in comparison. That said, the ZipGSX had a smaller chip count which also made it run cooler than the TWGS. And apparently the ZipGSX can't be overclocked as much as the TWGS, so for those who want to run the fastest still look to that board.

If even that... Because the system still has a built-in bottleneck of the 1MHz Mega-II core, which includes the graphics chipset and its RAM...

It's like a Mario, but with a funny accent. Most people pronounce a lot of words incorrectly. Especially in the USA, where they claim to speak "English"... but being an actual, real life, living, breathing native Englishman, living in England, and speaking actual English, I refute that claim! 😁

@@another3997/r/wooosh

ROM 3 doesn't, but the 0 and 1 do, and were designed to.

I'm pretty sure both of those are impossible. The CPU architecture is entirely different from the Macintosh line, and it's probably hard coded in how it's graphics work.

GSOS was an entirely different operating system than "System 6" for the Macintosh. So, to get System 7 to run on it, someone would have to re-write the Macintosh operating system to run on the IIgs hardware. In reality, GSOS already had many features of Macintosh System 7 either built-in or available as 3rd party software. (Truetype fonts and a "Multifinder"-like experience being examples of the latter)

yes some one makes a tap to get VGA out of it..

There was a VGA card for the IIgs called the Second Sight, which are pretty rare. Using an Oak VGA chipset, it provided both native graphics modes and enhanced VGA ones. Search for "iigs second sight" and you can find the manual for it.

1. It's been done, but... 2. A Mac on a card, basically.

Everything has it's day, the Apple II included. But that day was gone, things change, technology moves on, and some things just get left behind. At the time, Motorola's 68K processor line simply promised a more stable future than the WDC 65xxx series.

@@another3997 I know. I used to be a mega 68K Mac developer back in the day.

I'm not particularly bothered about Apple products in general, but that seems like a rather extreme attitude, considering it happened 30 years ago. Apple aren't the only company to sideline a portion of their customers in order to move things in another direction. Some people dispute the idea it was even Apple's fault, but rather than the chip manufacturer, WDC. After all, despite many promises, Motorola later struggled with 68K development, and both they and IBM eventually hit a wall with PowerPC. In both cases Apple were forced to change platform to survive.

@@another3997 It wasn't because of the speed of the machine and the reduced speed of the CPU, it's because of the way *Apple* treated their customers. They promised an upgrade, demoed it, then said "we're not releasing it, you'll just have to buy a Mac". This was actually the last good product Apple made. Macs have always been utter garbage. Their other products are cheaply made crap that they charge a premium for. When Apple's most powerful machine is less than 1/5th as powerful as my computer while costing more than twice as much, yeah it's lolworthy.