• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

68060 in Mac

Jockelill

Well-known member
We already have this on old Macs, either via PDS, NuBus or onboard Ethernet. Not sure why you need to put that on an accelerator or what that achieves?


There are several period NuBus video cards which offer high resolution support. Some are not easy to come by, look out in trading post for them.

As for most of what you are looking to experiment with, may I suggest trying something that isn't a compact Mac? Something with NuBus or PCI slots is going to give you much more flexibility/power/expandability.
It was mostly for the LC, here is only one PDS and getting hold of just a sonnet presto can be very difficult and getting an presto plus with Ethernet and 040 is very rare. Then yes of course, why not just buy a 475 or later ppc, but for the computers I listed I feel it would make sense. Of course one huge difference is, like @cheesestraw already pointed out, that the amiga never went past 68k whereas there is no shortage of Macintosh computers capable of running pretty much anything you’d throw at them.
 

Daniël

Well-known member
I think the main 68k Mac users interested in that many features, would be the ones rocking the workstations. And most of those have expandability that negates the need for a single solution card.

CPU Upgrade? PDS card, either old stock or a clone from Bolle when available.
More resolutions? NuBus cards (there's even work done on modern, HDMI capable NuBus cards!).
SCSI storage? Plethora of options (SCSI2SD, ArdSCSIno and derivatives, RaSCSI, ZuluSCSI, 80 pin hard drive adapters, IDE adapters, etc.).
Ethernet? NuBus cards, or a SCSI solution like RaSCSI and Scuznet.

Some of these options can be used for certain Compacts and LCs too.
I totally get the appeal of an FPGA, all-in-one card, and I don't mean to come across as hostile to your ideas, but I just don't think you can realistically get the same amount of heads together with the knowledge of this kind of work as the Apollo Team and other FPGA Amiga accelerator designers.

Maybe one day, a powerful, open source FPGA accelerator will be released that could be ported with less manpower to the Macintosh. Still, the peripherals that make it all-in-one would probably still require someone with Macintosh driver writing knowledge, which can be a scarcity.
 

egrath

Well-known member
so, wouldn’t it be possible to port the 68080 (now not the 60) in the same way to all the PDS-machines?
Your better bet would probably going the PiStorm route, which is essentially a combination of Raspberry Pi and a simple CPLD with the former one doing the actual heavylifting. In the current state, only regular 68000 DIP packaged can be replaced by it. Even if it's a project for Amigas, someone already has reported success in Macs:


/edit: Just as a sidenote, the PiStorm combined with a RPI 3A+ yields about 50x the performance of a 25 Mhz 68040 and 1700 times the performance of a regular 68k ....
 

MrFahrenheit

Well-known member
I’ve often wondered why anyone would spend $500+ on a CPU accelerator upgrade for an old Mac, when they could just buy a much faster Mac for less money.

Let’s consider the biggest target audience for upgrades: the SE/30

The SE/30 has the largest audience of “die-hard fans” of most of all models of Macintosh. You can upgrade the video to grayscale, install 68040 upgrades, and even have Ethernet. These upgrades, along with recapping analog board, and logic board, easily surpass $1000 US total.

Of the die-hard SE/30 fans, how many want an original “stock” SE/30 vs one of these super upgraded models? I would venture a guess it’s 10-20% of owners, tops.

In reading and talking with SE/30 owners, they seem to be the most die-hard beige Mac fans out there. Fans of pizza box Macs generally wouldn’t ever spend $1000 to upgrade / improve a Mac LC. I would venture a guess that the percent of pizza box Mac’s that want to spend more than $200 to make their Mac faster/better is single-digit percents.

Kay Koba makes a board that allows overclocking of the LC475/575 to 50mhz, and sometimes beyond, for under $100. VRAM upgrades for such a project, combined with a 40mhz full 68040 CPU, and a Spicy O’Clock would set someone back around $200. That’s relatively cheap in comparison to SE/30 upgrade costs.

I would estimate the number of people having done this upgrade is less than 50, worldwide. Some people want to push the 68040 to the fastest possible speeds. I applaud them. Some people ask “but what’s the point?” A 7200/90 is available for under $100, and if you really want to run stuff fast a G3 or G4 or G5 can be had usually around $100-150 tops. Keep the 475/575 the same if you own one, and just pickup a secondary faster Mac.

I’ve learned a lot doing the VRAM that I made: the market is so niche it’s almost not worth it to pursue.

For people like Bolle, I think it’s the fun and challenge of reverse-engineering and reproducing accelerator projects, more than anything. Bolle is targeting what is probably the largest audience of die-hard Mac fans, and even then his offerings don’t sell-out instantly.

At the end of the day, isn’t a PDS or processor upgrade that runs a faster processor in FPGA emulation no different than just sticking a Raspberry Pi inside the case and running an emulator?
 

Daniël

Well-known member
At the end of the day, isn’t a PDS or processor upgrade that runs a faster processor in FPGA emulation no different than just sticking a Raspberry Pi inside the case and running an emulator?

That's where things get philosophical. One could argue that an FPGA is more "authentic", because it can be programmed to be completely the same as a legitimate 68k chip. Of course, to get an 060 core working, you'd have to modify it to have the missing features, at which point it's no longer resembling a chip Motorola made 100%. An 080 by default isn't.

It can also be considered more "authentic" because the rest of the original hardware still does what it should. But what if the FPGA accelerator comes with its own RAM? Its own graphics? Its own storage? At that point, less and less of the original hardware is even used, merely becoming a host for the upgrade, one small step away from entirely replacing the original hardware.

And then, is something like PiStorm any less authentic? Or, indeed, just replacing the hardware entirely with emulation on a different architecture? All of this is philosophical ramblings, that you could fret over if you really wanted :)
 

Jockelill

Well-known member
I mean ideally a working 68060 (since daystar almost made one, it almost exists, they probably at least made a few working prototypes) like originally suggested in the thread would be very nice, but reading through the whole thread I sort of came to the conclusion that the community does not believe that to be possible. Since this amiga accelerator does exits, is built with modern parts and offers a lot of features, I thought it would at least be interesting to explore the idea.

Today the 68060 is also getting rare (they are also not made any more), sourcing a genuine 68040 is also somewhat difficult and getting hold of original accelerator can be very expensive. Some of these definitely goes in the 300-500range and even if Bolles rebuilds are really beautiful they still have the part sourcing issue.

All points made here are very clear to me, and it’s quite clear that such a project would likely not happen in the Mac community.

About adding features, the Sonnet Presto Plus added both new processor (68040RC33), more ram (32mb) and a Ethernet port, adding a micro SD to that would not too much more would it;)?

From a financial perspective any modern accelerator makes little sense, but so does the whole hobby doesn’t it :)?

if we just want to explore MacOS, sheepshaver runs it great on the M1/M2 macs, but after playing with tons of different emulators and virtual machines, a real old computer is simply much more fun☺️
 

Unknown_K

Well-known member
Back when I started collecting there were tons of cards and accelerators around and they were cheap so maxing out machines was fun and not that expensive except for a few super rare items.

These days those same cards and upgrades are much harder to find and many times as expensive. I like how Bolle reversed engineered some accelerators and sold them. While it might take a while, I think he pretty much sold them all so there is a segment of the hobby that will spend some cash for them. I suspect somehwere down the road there will be new motherboards made of current parts for some of the compacts and Quadras using non 68k CPUs under emulation where the end user will not know the difference outside of crazy speed kind of like Amiga offerings.

I personally like the old hardware but since prices have gone up so much, I rarely add anything to my 68K mac collection (not that I need to anyway). Most of what keeps me collecting these days is old PC hardware at the end of its value curve.
 

Phipli

Well-known member
Your better bet would probably going the PiStorm route, which is essentially a combination of Raspberry Pi and a simple CPLD with the former one doing the actual heavylifting. In the current state, only regular 68000 DIP packaged can be replaced by it. Even if it's a project for Amigas, someone already has reported success in Macs:


/edit: Just as a sidenote, the PiStorm combined with a RPI 3A+ yields about 50x the performance of a 25 Mhz 68040 and 1700 times the performance of a regular 68k ....
Someone asked this guy what it benchmarked at and he said 1.7x stock.

Some of the Amiga upgrades that drop in the CPU socket have been proposed for mac. I got a little frustrated because some people (I suspect not even the dev) were encouraging Mac people on facebook to fund the crowd funding thingy. They were talking about all the RAM and hundreds of % speedup. I belive that some compatibility is possible, but without likely re-writing parts of the ROM, making low level OS patches and and reigning in the expectations, this is going to take serious work. I feel that it was so unlikely to happen (I really hope it does!) that taking these people's money, or at least persuading people that are unlikely to benefit to fund a project they wanted to happen, saying it would give you a 512MB Mac Plus was... immoral.

As it stands, dropping in a faster 68000 replacement in a compact mac is going to cause the issues that the old upgrades struggled with back in the day - unlike the Amiga (I might be wrong - I've never owned one), the mac uses the CPU to control some timing critical hardware while turning off interrupts. These include the floppy drives (the cpu has control, directly, over spindle speed and head movement I believe), serial ports and sound at least. I don't know of a 68000 era mac processor upgrade that doesn't mess up sound!

This means that to get one working we need to catch ROM calls to functions related to these hardware devices and carefully make sure everything happens at the right times, likely slowing to stock speeds while doing it. The people with these skills are few and far between in the MacWorld sadly. The professionals seem to have moved on and even some of the documentation is hard to find. Another challenge with RAM is the 68000 Macs use the motherboard RAM as a framebuffer (two?) and the video circuit uses DMA. This would need to be considered. Also, the stock ROM (except the Portable?) is in the address map straight after the RAM, and you need continuous RAM on these machines. An MMU could solve this, but again, I don't have a clue where you'd start with that.

Getting a 68060 working is probably closer to possible, but nobody has one working that I know of. That 630 comment is really interesting - absolutely worth having this discussion just to learn that. I'd like to have a go one day, but might never learn enough to know how, be smart enough, or find the time.

I suspect all three. Plus my cat keeps sitting on me and I feel guilty pushing them off to do project work.

I imaging once you'd gone through all manner of horrors patching the ROM, you'd set up traps to capture the unimplemented Opcodes and load them in an extension as soon as possible. Then it would effectively live patch any software. Hopefully the OS doesn't call any early on or... more pain.

Since we're talking about it, what I want to see is someone bodge a PPC750 onto a Daystar 601 ;)

Its all 601 bus! 100MHz G3 in a IIci?! An inefficient waste of time? I didn't realise you knew me so well ;) Although thankfully I'm not crippled by a bus...

Lastly, we've probably been a little too negative and I just want to say, if you start trying to implement the mods needed, I may not be much use but I would do everything I can to help. I suspect progress would be the biggest motivator. As things are, I don't expect to see it happen anytime soon.
 
Last edited:

LaPorta

Well-known member
As for the SE/30 fans, I have an analogous story. It isn’t quite the same, but might follow a similar path. World War II aircraft were a dime a dozen in the decades following the war. Immediately after the war, most were either scrapped or junked. Others, some people used for work. Some modded them for air races. Others became personal planes.

Fast-forward to the 80s and 90s. People began to see value in not only preserving what remained of these aircraft, but also returning them to their original wartime condition. Mods were removed, original guns and turrets reinstalled. Originality and authenticity was strived for.

Im not saying this will happen, but it could in the future. Of course, there will always be a place for the machine with a cool history that was modded to serve a purpose (like the story of that metal-encased NASA IIfx on here). But those remaining with their near original logic boards, etc, may one day be more valuable than the modded ones.
 

Jockelill

Well-known member
I mean ideally a working 68060 (since daystar almost made one, it almost exists, they probably at least made a few working prototypes) like originally suggested in the thread would be very nice, but reading through the whole thread I sort of came to the conclusion that the community does not believe that to be possible. Since this amiga accelerator does exits, is built with modern parts and offers a lot of features, I thought it would at least be interesting to explore the idea.

Today the 68060 is also getting rare (they are also not made any more), sourcing a genuine 68040 is also somewhat difficult and getting hold of original accelerator can be very expensive. Some of these definitely goes in the 300-500range and even if Bolles rebuilds are really beautiful they still have the part sourcing issue.

All points made here are very clear to , and it’s quite clear that such a project would likely not happen in the Mac community.

About adding features, the Sonnet Presto Plus added both new processor (68040RC33), more ram (32mb) and a Ethernet port, adding a micro SD to that would not too much more would it;)?

From a financial perspective any modern accelerator makes little sense, but so does the whole hobby doesn’t it :)?

if we just want to explore MacOS, sheepshaver runs it great on the M1/M2 macs, but after playing with tons of different emulators and virtual machines, a real old computer is simply much more fun☺️
 

cloverskull

Well-known member
The 68060 has benefits beyond merely speed, the biggest benefit IMO being that they run a heck of a lot cooler! For that reason alone I would love to run a true 060. :)
 

Unknown_K

Well-known member
The 68060 has benefits beyond merely speed, the biggest benefit IMO being that they run a heck of a lot cooler! For that reason alone I would love to run a true 060. :)
A 68040/33 has max power of 6.6W, quit bitching about heat. Do you run a 300W TDP I9 at home?
 

cheesestraws

Well-known member
The 68060 has benefits beyond merely speed, the biggest benefit IMO being that they run a heck of a lot cooler! For that reason alone I would love to run a true 060. :)

Yeah, this is one of the projects that I'd love to have the time and expertise to do, but unfortunately I have neither - not sure anyone else has the expertise either.

quit bitching about heat

This ... seems like a bit of an overreaction? Quite a lot of an overreaction, in fact.
 

Unknown_K

Well-known member
Yea, a bit of an overreaction sorry.

I just don't think they run that hot unless you are overclocking them or using a chip that has been relabeled for a higher speed. A simple heatsink with minimal airflow is good enough to keep you from getting burned.
 

cheesestraws

Well-known member
More heat is more thermal stress when the machine is powered on or off. Thermal stress shortens the life of solder joints, connectors, etc, by causing movement in them.

Lower change of temperature is good in something you want to keep running long term.
 

Phipli

Well-known member
More heat is more thermal stress when the machine is powered on or off. Thermal stress shortens the life of solder joints, connectors, etc, by causing movement in them.

Lower change of temperature is good in something you want to keep running long term.
Better for keeping the teapot warm though.
 

NJRoadfan

Well-known member
The 25Mhz 68040 in my Amiga 4000 has a heatsink on it and gets quite toasty. The ZERO airflow in the case is where the problem lays. Scary to think that people run Cyberstorm PPC cards in these things. That card has both a 68060 and a PowerPC 604e on it!
 

joshc

Well-known member
Depending on the version of 040 they do run hot especially at higher speeds. My D50D mask full 040/25 runs toasty at 33 so it needs a heatsink otherwise causes my 475 to crash/stop running. Since I'm moving to running the 475 at 45mhz, I am switching to a E31F chip instead which should run better at higher speeds.
 
Top