@Beans

Quote:

considering the rather poor start, (in the 8088/8086 family)

8088 was simple, but simply genial CPU. Ideal price (8 bit peripheals) and performance (16 bit internals with support for up to 1 MB RAM) was best suited for PC of the early 1980s.

68000 on the other hand would then hardly fit into lowcost (under 2000 USD) design. Note 68008 was not released until 1982.

@kamelit0

Quote:

It failed because IBM was not able to deliver a power efficient G5 for laptop so Apple ditch them.

Even low-power G4 couldn´t compete with early Pentium M. Steve Jobs was marketing genius, else Apple wouldn´t survive with such inferior (performance-wise) products.

What ever happened to Dec alpha chips

_________________
A1200 ACA1232 128MB Indivison MkIICr
Elbox empty Power Tower
RPi3 Emulating C64 ZX Atari PS BBC
Wii with Amiga emulation
Vampire v4 SA

@pavlor

Mac OS X was improving beyond what Windows XP could provide (for home and creative users anyway). G4 laptops were good enough and the best cost to performance ratios ever for Apple laptops! iBook G4s were great for their time.

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

@BigD

Quote:

G4 laptops were good enough and the best cost to performance ratios ever for Apple laptops! iBook G4s were great for their time.

Intel based Apple laptops had (much) better price/performance ratio when runing native applications.

Apple´s story shows real performance is not paramount, if you know, how to sell your computers. This reminds me of Vampire benchmarks emphasising strong points of their design (relative to 68060 - eg. fast RAM), but omitting weak points (FPU). I even saw general comparison of V4 to 300 MHz 68060, which is valid, but only for selected benchmark (video playback using advantage of fast RAM), not as measure of general performance. However, most important is not real performance (be it G4, G5, Vampire, or Tabor), but ability to persuade customers your product is exactly what they want. Apple was stellar at this and people behind Vampire seem to do right job too.

@pavlor

When you compare a processor design from 1999 and one from 2006, Intel better had a better chip till then. The G3 and G4 in their time were generally better chips then anything Intel or AMD had. The G5 was quite powerful when it came out, but of course Intel had something more efficient 3 years later.

The whole performance/watt thing was just an argument for Jobs for what he wanted to do for a long time, transition to x86. Performance and efficiency were not his real concerns.

@WolfpackN64

Quote:

The G3 and G4 in their time were generally better chips then anything Intel or AMD had.

Pentium M was released in 2003. It was competitive even to G5:

Pentium M 755 2 GHz (benchmarked in June 2004):
1541 SpecInt2000
1088 SpecFp2000

970 2.2 GHz (benchmarked in November 2004):
1040 SpecInt2000
1241 SpecFp2000

@pavlor

I'd like to see a 2Ghz G4 next to that. Since the G4 was faster in some ways as the G5 clock per clock (except in floating point), it goes how far to show the G4 was a very performant chip

@WolfpackN64

G4 worship became like a religion!

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

@BigD

Oh God, who even makes that kind of artwork for a processor?

@WolfpackN64

Quote:

I'd like to see a 2Ghz G4 next to that.

Rather not, I could fall from my chair... laughting.


To be more serious, there were no G4 above 1.666 GHz. Apple used 1.5 GHz G4 in 2004.

@pavlor

Newertech had a dual 1.8 Ghz CPU card for Mac's and Gigadesign had a single 2.0 Ghz CPU card. Both were G4 CPU's.

There were even benchmarks of the 2Ghz card.
http://barefeats.com/g4up.html

Last edited by WolfpackN64 on 07-Aug-2017 at 05:17 PM.

@WolfpackN64

Sorry, I meant CPU used laptops. However, 2 GHz G4 couldn´t be much faster than 2.2 GHz G5.

@pavlor

Quote:

8088 was simple, but simply genial CPU. Ideal price (8 bit peripheals) and performance (16 bit internals with support for up to 1 MB RAM) was best suited for PC of the early 1980s. 68000 on the other hand would then hardly fit into lowcost...

Yeah, I started out with 6800 and 6809 cpus, and didn't migrate to the 68000 until the 80186 was introduced.
My jaundiced opinion about Intel continued through the 80286 as well, which was a rather underwhelming processor.

And yes, the 68000 was not suited to $200 systems. Our own 68K systems started at $999, and while base Atari STs might have been slightly less, they didn't come with hard drives.

Then again, I wouldn't characterize early 8088/8086 systems as cheap either (unless you were looking at something like a Tandy 1000).

And at 4.77MHz, they benchmarked really poorly.

All you need to do is clock a 6502, 6800, or 6809 to 2 MHz, and the performance advantages of the 8088 evaporate.
The basic architecture (along with Zilog's), runs fast, but doesn't work that efficiently.

Although, I did rather like the NEC V20 and V30 cpus of that period. Cheap and faster than Intel alternatives.

As I've said before (many times too often), my basic objections to X86 were nullified once the '386 was introduced (and X86 continued to improve through the '486 to the Pentium). '686/Pentium architecture was the final nail in the coffin for the 68K.

I shifted to X86 at about the time Socket7 cpus were being introduced. That was a fun period (I still have a few AMD K6-2+ and K6-III+ cpus from my last Socket 7 systems).

_________________
Amiga! "Our appeal has become more selective"

@Beans

Quote:

And yes, the 68000 was not suited to $200 systems.

Well, $2000 (not 200) was considered "lowend" for a business computer in the early 1980s. 8088 had great advantage over 8 bit CPUs by support for far more RAM. By 1981, bussiness computers reached memory limit of 8 bit CPUs (usualy 64 kB).

@pavlor

Probably not no, but remember, It was suddenly all about performance/watt.

@pavlor

Quote:

...64 kB

Without an MMU, yeah, and even with the addressing limitation had some consequences.

But does everybody remember what memory used to cost?

Investing in a full 512 or 640K often cost several hundred dollars.
These days we take cheap memory for granted, but when these systems were new there was a serious supply problem.

While the 68000 addresses 16MB (and please, don't start quoting the architecture's maximums), few systems ever supported that 'lofty' amount.

Hey, remember "640K ought to be enough for anybody". A full TEN TIMES what 8 bit cpus could address!

@WolfpackN64

Quote:

It was suddenly all about performance/watt

@pavlor

Quote:

Note 68008 was not released until 1982.

The 68008 is Motorola's answer to the 8088.
Easy use of 8 bit peripherals, cheaper designs for memory, still it really performs quite a bit worse than the standard 68000.

I've never used one.

I don't remember that being an important consideration, at least not until recently.

Last edited by Beans on 07-Aug-2017 at 06:22 PM.
Last edited by Beans on 07-Aug-2017 at 06:18 PM.
Last edited by Beans on 07-Aug-2017 at 06:16 PM.

_________________
Amiga! "Our appeal has become more selective"

@Beans

Quote:

But does everybody remember what memory used to cost?

Mid 1982:
64 kB memory expansion for IBM PC - 449 USD
128 kB - 679 USD
192 kB - 879 USD
(bulk of the price is the card itself)

@Beans

Have you seen the conference where Jobs announced them going to Intel? Performance/watt was the buzzword of the day. Of course the Core 2 Duo was more efficient then the G5...

@WolfpackN64

Jobs certainly liked buzzwords, but the idea that the G5 is significantly less efficient than X64 cpus (which can draw 125 watts or more) is a bit ridiculous.

ARM cpus and embedded PPCs don't draw as much as the 'lower wattage' X86 cpus.
And Core2? Hardly the model of efficiency.

_________________
Amiga! "Our appeal has become more selective"