Commodore 256 under construction


#41

I think running Linux and NetBSD natively is better than trying to wedge it into a DOS compatible card.


#42

I dimly remember that the the ppc port (mkLinux) wasn’t then compatible with the PowerMac 6100 (and siblings)


#43

Regarding the venerable 68K: the last iteration of the 68K family was the 68040, which maxed out at under 50Mhz. it was largely succeeded by the PowerPC 600 series, which was one of the few times where you had IBM engineers in the same room as Apple engineers. :slight_smile: The Power PC sort-of lives on as IBM’s POWER8 and POWER9 based servers and workstations. ( [RedactedCo] has a couple pairs of P550 servers rocking the Power6 processor, and we just got in a pair of S814s running POWER8 processors (which are ludicrously over-spec for the load we have on them.)


#44

unfortunately mkLinux was not Linux at all. It emulated the Linux system calls but it was a completely different kernel. It was more closely related to Darwin/OSX than Linux in terms of kernel and architecture.

I mainly ran Linux on PowerPC PReP boards, much better supported than the closed Mac ecosystem. There was some Linux support for the Power Macintosh 7200, but it didn’t really go much earlier than that (as that series was similar to an early open standard CHRP)

Wow I’m full of weird computer trivia today. Feeling my age too.


#45

The 68000 processor is sometimes described as 16-bit (it has a 16-bit external data bus and ALU), sometimes as 32-bit (32-bit internal data bus and registers).

The Sinclair QL used the closely related 68008, which only had an 8-bit external data bus, making it slower than most computers of the same generation.


#46

The 68060 came later and ran 50Mhz to 75Mhz.

After the 68060 a vast family of embedded 68K CPUs came out with things like on die RAM, but that was a fair bit later.

The 68040 was the last 68K CPU I did any real programming on though. In an embedded crypto …when crypto wasn’t currency but DES. It was pretty fast, but again pretty fast was encrypting a T1 (that’s 1.5Mbits/second)…

It was very competitive with the 386s/486s of the era…but then didn’t keep up.

I think of the 68k could have kept up of Motorola could have afforded the R&D, but with the kind of volume Inte was shipping they could charge less and still afford far more R&D. Whatever was in the volume position was going to have the most R&D. One of those virtuous circle things that really isn’t all that virtuous…


#47

We can slice it finer if you want, but having 32 bit registers and a full set of instructions operating on them makes it a 32 bit computer even if the ALU is bitslice. The ‘16 bittness’ of the ALU matters very little. And the address and external data busses don’t matter much or the PC would be considered an 8 bit machine.

The 68K family went to the 68060. Maybe they weren’t used outside of Motorola at the time, I don’t know because I was inside the machine then.


#48

currency is but a faddish sideline of crypto.


#49

Those figures are only for the US? My sources suggest that over 3.5 million ZX Spectrum 16k/48k/+ were sold. Other than RAM, those models were functionally identical (The + had a better keyboard). I have no idea how many of the Timex and unofficial eastern bloc clones were sold. I expect that worldwide the C64 beat the ZX Spectrum (and clones) in the 8 bit wars though.

Edit: Wikipedia suggests 5 million units sold, but with no source.


#50

There were no 32-bit multiply or divide instructions in the 68000, for what it’s worth- these were only available from the 68020 on.


#51

I sit corrected. Thank you!

I do find it amusing in an ironic way that we are currently using ‘low end’ processors that were unheard of back (1000 Mhz?! UNPOSSIBLE!) then to emuate the entire hardware stack in software (cpu and all the support asics), and in some cases have to throttle the virtual machine down because it runs too fast.


#52

I think that’s the big blob of “other” in around 1982, there. The Ars source article is here.


#53

It doesn’t fit with the figures. The issue 3 motherboard was put in Spectrums sold from 1983 at the earliest, which was the most common board sold (3 million).

I don’t think the Ars Technica numbers are wrong for the US, but they don’t seem to be right for the rest of the world.


#54

You had ones? Luxury!


#55

Well, yes, the rest of the world is irrelevant, isn’t it.


#56

This topic was automatically closed after 5 days. New replies are no longer allowed.