Why did 68k or powerpc fail?

Why did 68k or powerpc fail?

Other urls found in this thread:

mil-embedded.com/pdfs/Freescale.Sep07.pdf
twitter.com/NSFWRedditVideo

68k failed because it used old design, the 68060 partly remedied that but sacrificed too much for backwards compatability.

PPC failed simply because it couldn't compete with x86 in regards to software library and wasn't exactly the easiest architecture to develop for.

I still miss the old 68k days and assembly straight on the hardware, but those are long gone.

They didn't fail, they were fairly successful back then.

Just got obsoleted is all

Install Gent.oo.

doesn't modern x86 also keep a lot of legacy stuff?

Not to the same extent. The really ancient stuff like x87 and so on is emulated in the hardware.

>and wasn't exactly the easiest architecture to develop for.
Are you kidding? Assembly was much cleaner on PPC than on quirky x86. Of course the vast majority of people don't write assembly so this is a moot point.

The real reason was your first point: Windows compatibility carried x86 through the RISC assault. And before someone replies "hurr durr windows ran on risc", I mean compatibility with the massive Wintel software library without slow emulation.

People already knew how to program for the "quirky x86", they didn't fully know what to make of the PPC.

Also: 68K "failed" simply because Apple, IBM, and Motorola believed RISC would give them a large enough price/performance advantage to unseat Intel. And all three halted their other work to focus on PPC, with the exception of Apple's investments in ARM (we're talking Newton days here, not iOS).

PPC did hold a significant price/performance advantage. And if Moore's Law had failed around the time of the 604e or G3 it might indeed have unseated x86. But Moore's Law kept going, and most people aren't going to ditch their software library when they can wait 6-9 months for similar performance from x86.

>People already knew how to program for the "quirky x86", they didn't fully know what to make of the PPC.
As someone who programmed professionally during this time period: if you could write x86, PPC was cake. PPC is a very clean and logical ISA that's easy to write optimal code for.

But again, the vast majority of code was not written in assembler for either. If you were a Windows 9x developer back then the difficulty was in learning Apple's OS and APIs. C/C++ compiled the same on either platform.

Apple ditched PowerPC because they got better price:performance from Intel. The official reason was that "Intel had a better roadmap" which was also true. After they dropped PowerPC it fell out of mainstream use except for maybe some embedded stuff.

For the same reason that Betamax failed.

So x86 had the porn?

Think it was they couldn't get a G5 chip in a Mac laptop (there was a prototype). Mobile roadmap and R&D problems too.

> Back in the G4 days, 167Mhz bus...

I'd like to dedicate a memory to two architectures:

HP's PA-RISC that was successful but started running out of steam and was succeeded by Itanium,

and DEC's Alpha which was really great, but when Compaq acquired DEC, they decided that Itanium had a better future and killed off Alpha. What a colossal mistake.

Fuck Itanium.

Thank god AMD killed Itanium.

this, IBM could not deliver a G5 for powerbooks and ibooks, years after they introduced it to powermacs and imacs.

It lost priority when PPC was on the ascent.

Otherwise it was successful, also in embedded and in the Palm.

Worth noting that PowerPC went on to become OpenPOWER, and Power7 systems are worth keeping if you find them.

POWER and PowerPC are two separate architectures. PowerPC's basically only used in embedded devices now. For example, RAID cards, NASs, etc.

>HP's PA-RISC that was successful but started running out of steam and was succeeded by Itanium,
I heard that PA-RISC had a design misfeature in that the integer and floating point registers wer eunified and that led to register file contention which meant a serioius performance hit.

I am not sure this aspect was a user visible part of the ISA so it may be that a re-engineering would have worked while maintaining the ISA.

Mind you the ISA op code chart was horrifically complicated and made x86 look trivial in comparison.

PPC didn't fail, and they're still used in microcontroller and in embedded system, for both commercial airliner and military electronic suites (like Thales Mirador)

Also, a good read about PPC application in the military

mil-embedded.com/pdfs/Freescale.Sep07.pdf

The military is always on the back end of technological curve, man. Commercial development take place first before it is adapted for military application

Fuck we're still using coleron-tier processor in military satellites

You mean ppc64el? I thought they were bascially the same.

PowerPC's based on POWER