Generations of Personal Computers

[Kitsune]

I you were to list what the generations of Personal Computers are.
Obviously, a two to three gigahertz Quad Core modern processor is something completely different than a 4 megahertz 8088 processor.
How would people here define the various Person Computer generations?

[Drooling Iguana]

It's tough to define PC generations since there's a hell of a lot of overlap. It's not like consoles with a fairly strict 5-year lifecycle. I'm not sure exactly where I'd draw the lines, but major turning points would be the transition from the BASIC era (Apple II, Commodore 64, etc.) to the DOS era and the DOS era to the Windows 95+ era, with things like the introduction of 3D acceleration and ubiquitous Internet connectivity also being important events.

[Count Chocula]

I'd classify my first computer, a Radio Shack TRS-80 with a color TV monitor, 64kb RAM and rockin' cassette tape drive, as second generation. The Altair and Heathkit 8088 processor computers are those I'd classify as first-generation PCs.

[weemadando]

Well, I've had the following which represent a pretty decent coverage of the past 25 years of tech.

Apple IIe - probably a 3rd or very early 4th gen home PC.
386
486
Celeron 300
P3-800
P4-2.4
Dual Core E8400

[Count Chocula]

Don't forget the Amiga: that was a machine at least 5 years ahead of its time. I'd call it an early 3rd generation computer, with the original IBM PC as the "official" 3rd generation, with its 286 processor and DOS. The Apple II I'd call a 2nd generation unit, but the IIe was definitely on par with the IBM PC...without the software options. My freshman year CompSci computer, BTW, was a dog-slow Apple Lisa with an external 10MB Unix hard drive; my Engineering roommate had an IBM PC, for comparison. This was in 1985.

[Kitsune]

Well, kind of trying to figure out how to judge my computers

Desktops
8088 (Tandy 1000) - DOS
286-16 (AT Board) - DOS
386-40 (AT Board) - DOS
586-133 (AT Board) - Win 95
K6-350 (AT Board) - Win 98
Athlon 900 (ATX Board - Slot - died) - Win 98
Athlon 1.2 (ATX Board - Socket - died) - Win 98
K6-450 (ATX Board - Used Emergency Replacement) - Win 98SE
Sempron 2600 and an Athlon XP 2800 (Both ATX Board) - Win XP

Laptops:
386SX (Grey Screen) - DOS
Pentium-100 (Color S-VGA) - Win 95
Pentium-166 (Active Matrix) - Win 98
Pentium III - 850 - Win NT 4.0 / 98 SE / XP Pro
Pentium IV - 2 Ghz - XP Pro

I also built a friend a computer using a Centrino 1.3 and told the person that it only about a generation back and wondering if that is a good estimate?

[Uraniun235]

I don't think it's productive to try and define any sort of 'generation' for something as large as "personal computers" - it's been a gradual shift over the years. I think the best way is to consider the widespread adoption of certain technologies. The hard drive, CD-ROM drive, and sound card were all important steps towards what we consider a PC today. I'd also suggest that USB flash drives are what finally enabled us to start dropping the floppy drive, and CD burners allowed us to both create our own music CDs and to get away from the Zip drive. Wireless networking is also an interesting development, if absolutely infuriating to tussle with at times.

I also built a friend a computer using a Centrino 1.3 and told the person that it only about a generation back and wondering if that is a good estimate?

The only "Centrino 1.3" chip I'm seeing (by the way, Centrino refers to a combination of hardware - it means it has a certain wireless adapter, a certain chipset, and a certain type of CPU) is a Pentium-M that was released in 2004.

I'm really curious to know how you would define that as being "only about a generation back". From a purely Intel processor-oriented perspective, since Intel is currently manufacturing and selling the Core i7 series as their new product line, I would consider the Core 2 series to be "last generation". But even if we step back from Intel and take a broader look, I still think the last generation would be more like the later Pentium 4 and Athlon XP series chips, which would outperform that Pentium-M chip in nearly everything except power efficiency.


Really though, it's not at all descriptive to say "oh this chip's just last generation" or "two generations ago". It's a lot more descriptive to say that it's four years old, or that it would be totally outclassed by any CPU sold as new today.

[Glocksman]

My first computer was a Timex/Sinclair 1000 complete with 16k RAM pack and thermal printer.
It had the singular advantage of being really cheap for the time ($100 for all three pieces at Woolco's 1982 going out of business sale).

Black and white video and a Zilog Z80A running at 3.25Mhz.
I used it for really basic gaming and writing BASIC programs at home for my computer math class.

This site has a machine timeline that's interesting.
oldcomputers.net

[Seggybop]

But even if we step back from Intel and take a broader look, I still think the last generation would be more like the later Pentium 4 and Athlon XP series chips, which would outperform that Pentium-M chip in nearly everything except power efficiency.

This is not necessarily so; the Pentium M is generally at least 10% faster at the same clock speed as an Athlon XP, and more than twice as fast as a Pentium 4. Core Duo was basically the combination of two tweaked Pentium M cores, C2D being further incremental improvements from there.

It's also worth noting it's way faster than anything new using an Intel Atom without much worse power consumption, and cheaper due to the age.

[Starglider]

There was a very definite 'generational split' between 8-bit and 16-bit personal computers, and a fairly pronounced one between 16-bit and 32-bit personal computers. However everything after that has been fairly incremental. For example the 32-bit to 64-bit transition is taking something like 5 years with most people barely noticing that it's happening.

My computer ownership history (at least, the machines I got a lot of use out of);
ZX Spectrum 48k
Amiga 500
Apricot XEN-i
Amiga 1200
Windows 95 PC with AMD K5 PR166
Windows 98 PC with 450 MHz Pentium3
Windows 2000 PC with 1600 MHz Athlon (later upgraded to 2GHz)
Windows XP laptop with a mobile Athlon.

Currently I have a quad core Vista laptop (overclocked to 3 GHz) and a 8-core 2.5 GHz Xeon workstation running Vista and Debian. I'll be getting a Nehalem machine when the dual-processor chips come out, though I might wait for the 6-core ones.

[TimothyC]

My Parents would buy computers every 4 years or so:
1988: Zenith Z-157; 8088 @ 8MHz, 1 meg of ram, 20 meg hard drive [MS-DOS 3.30 to 5.0 in 91]
1992: Generic; 386DX @ 25MHz, 2 megs of ram (upgraded to 8 in 93), 80 meg hard drive (upgraded to 540 in 93) [WIN 3.0 to WIN3.11 in 94]
1996: Packard Bell; P120, 8 megs of ram, 1 gig hard drive (upgraded to 11 gigs in 1999) [WIN95, to WIN98SE in 99]
2000: Gateway; K7 @ 700MHz, 64 megs of ram, 10 gig hard drive [WIN98SE]
2004: Dell; P4 @ 2.66 GHz, 1 gig of ram, 100 gig hard drive [XP Home]
2004 (Laptop - Mine): Dell Insprion 5150; P4 Mobile @ 2.66, 256 megs of ram (to 512 in 2005), 40 gig hard drive [XP Home to Pro]
2009: Dell; Core 2 Quad, 4 gigs of ram, 250 gig hard drive, [XP Pro with Vista Ultimate on Disk]

Every single upgrade was a massive upgrade from the prior generation.

[Uraniun235]

But even if we step back from Intel and take a broader look, I still think the last generation would be more like the later Pentium 4 and Athlon XP series chips, which would outperform that Pentium-M chip in nearly everything except power efficiency.

This is not necessarily so; the Pentium M is generally at least 10% faster at the same clock speed as an Athlon XP, and more than twice as fast as a Pentium 4. Core Duo was basically the combination of two tweaked Pentium M cores, C2D being further incremental improvements from there.

It's also worth noting it's way faster than anything new using an Intel Atom without much worse power consumption, and cheaper due to the age.

I see your point, but at the same time it highlights the reason why looking at processors in terms of "generations" isn't very useful when you're trying to sell someone a computer - those Pentium-M chips aren't going to be running at anywhere near the same clockspeed, I'm pretty sure nearly any new Athlon XP or Pentium 4 chip released in 2004 is going to be significantly faster than a Pentium-M at only 1.3GHz... which, by the way, fits with my remark about power efficiency.

Are we talking about a laptop or a desktop here? I ask because I think I've heard of people sticking mobile processors in desktop motherboards before and you normally don't hear about people "building" laptops for other people.

[Kitsune]

I screwed up what I said earlier, I should have stated "Celeron" 1.3 Ghz processor not "Centrino"...Sorry!

It was a desktop machine built out of salvaged parts. Started as an E-Machine, replaced the power supply with something more robust (The old one was I think 100 or 125 Watt [Which also was dead] replaced with a 250 Watt), increased RAM to 512 (had 128), put a 30 gig hard drive in, and replaced the CD ROM with a read-write CD Rom.

Might have also crewed up in that it might have been a 1.2 Ghz processor.

I was just trying to make all my descriptions as accurate as possible, making sure that I told them the right information.

[Jade Falcon]

My first computer was a Timex/Sinclair 1000 complete with 16k RAM pack and thermal printer.
It had the singular advantage of being really cheap for the time ($100 for all three pieces at Woolco's 1982 going out of business sale).

Black and white video and a Zilog Z80A running at 3.25Mhz.
I used it for really basic gaming and writing BASIC programs at home for my computer math class.

This site has a machine timeline that's interesting.
oldcomputers.net

Ah, that was what was known over here as the Sinclair ZX-81. Did you have the Sinclair printer or the alternative Alphacom printer?

Myself...I have had.

Commodore VIC 20 3.5k RAM
ZX Spectrum 128k
ZX Spectrum 128k +2
Atari ST
Escom 486 sx25
Upgraded later ot 486 DX/66
Cyrix 166mhz
AMD K62 350
AMD 1.2ghz
AMD 2000
IBM Quad Core Q6600

[The Kernel]

Here's my history of my PCs as long as this is turning into a computer museum.

Toshiba 80286 laptop

Macintosh Classic
8 MHz Motorola 68000
1MB RAM
512KB ROM

Macintosh Color Classic
16 MHz Motorola 68030
4MB RAM

Intel 486 DX2 66MHz
8MB RAM

Intel Pentium 75MHz
16MB RAM
Speedstar Pro

Intel Pentium Pro 200MHz
64MB RAM
NV1 Diamond Edge 3D

Intel Pentium II "Deschutes" 233MHz
64MB RAM
ATI All-in-Wonder

Intel Pentium III "Coppermine" 733MHz
128MB RAM
3dfx Voodoo 2

AMD Athlon 1.2 GHz
512MB RAM
GeForce 2 GTS

Pentium 4 "Northwood" 2.0GHz
1GB RAM
GeForce 4 4600

Athlon 64 3500+
2GB RAM
GeForce 6800GT

Core 2 Duo E6400
2GB RAM
GeForce 6800GT

Core 2 Quad Q600
2GB RAM
GeForce 8800GT

Core i7 920
6GB RAM
GeForce 8800GT

[Drooling Iguana]

Listing all the computers I've owned would be a bit difficult since technically I'm still using the same computer I used in 1992. Every component has been replaced several times, of course, but never all at once.

[Kitsune]

Didn't you have to effectively switch computers when you went from AT to ATX?

[Drooling Iguana]

Didn't you have to effectively switch computers when you went from AT to ATX?

Kept the optical drive. Might've kept the video card, too, I can't remember.

[Ariphaos]

I would go by memory addressing schemes - it's a harder limit and even though that line is frequently blurred in various implementations, it essentially declares what the system is theoretically capable of, as opposed to simply lacking in horsepower or available instructions. You can actually go a lot further in terms of what generation constitutes what advancement.

First Generation would be raw 16-bit addressing and 'cheating' schemes, 16+8 or whatever, from 1977 with the release of the first PC to 1994 when Commodore finally died - or more realistically in 1992 with the last Apple II.
Second Gen would be heralded by the i386 series and other genuinely 32 bit systems. From 1986 with the release of the i386 to 1999 when Intel discontinued the Pentium.
Third Generation would begin in 1995 with the Pentium Pro and all sorts of advanced tricks in preparation for, amongst other things, PAE and similar cheating addressing schemes. Ends in 2008 when Intel discontinues all desktop processors not supporting x86-64.
Fourth Generation would be genuine 64 bit architectures - x86-64. In 2006 Apple dropped the PowerPC for Intel processors. From 2003 to present.

Fifth generation is almost certainly going the development of a truly usable NVRAM technology that becomes ubiquitous such as MRAM or racetrack memory.

[Kitsune]

Well, in several switches, some components such as memory, CD Roms, and floppy drives were shared.

My second through fourth computer shared the same AT Case
Fifth and sixth shared the same ATX case
There was a bug in the machine so that is why I went with an all new (used) K6 machine
I kind of consider it a new computer anytime I replace the motherboard

[Starglider]

I would go by memory addressing schemes - it's a harder limit and even though that line is frequently blurred in various implementations, it essentially declares what the system is theoretically capable of, as opposed to simply lacking in horsepower or available instructions.

As such it's pretty much irrelevant for practical purposes. Very few users have ever run into addressing limitations. If you're going to pick an arbitrary technical specification, the feature size or transistor count of the processor (and GPU where applicable) would have a much better correlation with actual capabilities available to the user. Computer generations used to be numbered this way (first through fourth where first is vacuum tubes, second is transistors, third is ICs and fourth is VLSI), but that scheme kind of fell out of use as litho progress got more incremental. Technically since the threshold for VLSI was about 1,000 (enabling the first single-chip microprocessors; the Z80, 6502 and 4004 had a few thousand transistors each), you could say that fifth generation was the 1,000,000 transistor threshold (reached with the 486) and that we are just making the transition to 1,000,000,000 transistor sixth generation now.

Fifth generation is almost certainly going the development of a truly usable NVRAM technology that becomes ubiquitous such as MRAM or racetrack memory.

That won't make all that much difference to most users. Most of those technologies are moderately superior replacements for current Flash technology. Even the rise of Flash and the death of the HD is mostly relevant for power and startup time, it doesn't make a huge amount of difference to the processing most users do.

[Stark]

Well, in several switches, some components such as memory, CD Roms, and floppy drives were shared.

My second through fourth computer shared the same AT Case
Fifth and sixth shared the same ATX case
There was a bug in the machine so that is why I went with an all new (used) K6 machine
I kind of consider it a new computer anytime I replace the motherboard

'A bug in the machine'? Oh dear.

[Glocksman]

Ah, that was what was known over here as the Sinclair ZX-81. Did you have the Sinclair printer or the alternative Alphacom printer?

It was the Timex printer that looked like an oversized receipt printer and used rolls of thermal paper and after looking it up, it turns out to be a rebranded Alphacom 32.
I guess Timex didn't want to license Sinclair's printer.
Interestingly enough, most of the hobbyist magazines I bought at the time referred to it as a 'ZX-81' in many of the articles.
That and the misspelling of the word 'color' , told me that the US publishers didn't even bother editing them for a US audience.

I learned early on to rubberband the printer interface/RAM pack combo to the unit after losing a 1/2 hours work of keying in a program by jostling the table.

[Jade Falcon]

Ah, that was what was known over here as the Sinclair ZX-81. Did you have the Sinclair printer or the alternative Alphacom printer?

It was the Timex printer that looked like an oversized receipt printer and used rolls of thermal paper and after looking it up, it turns out to be a rebranded Alphacom 32.
I guess Timex didn't want to license Sinclair's printer.
Interestingly enough, most of the hobbyist magazines I bought at the time referred to it as a 'ZX-81' in many of the articles.
That and the misspelling of the word 'color' , told me that the US publishers didn't even bother editing them for a US audience.

I learned early on to rubberband the printer interface/RAM pack combo to the unit after losing a 1/2 hours work of keying in a program by jostling the table.

The receipt type printers were rather funny weren't they. Ah, that's when a Star LC-10 dot matrix seemed like hi tech stuff.

Oh, and you refer of course to what was known over here as 'Ram Pack wobble'. Then of course for Sinclair/Timex computers you had to buy a joystick interface. Over here there were three prime competitors in the Spectrum days. There was the Protek, Kempston and the Sinclair. Sinclair tried to standardise with their 'own brand' but Kemptson was the main choice. They Sinclair interface also had a cartridge slot which accepted cartridge games, though as far as I know the only cartridge games were direct ports of the Horace series, Hungry Horace (basically a Pac Man clone), Horace goes Skiing and one other game which I think was a Frogger type.

[Starglider]

Ah, that's when a Star LC-10 dot matrix seemed like hi tech stuff.

Hah, I had a Star LC-10 Colour, connected to our A500. I remember having to embed all the styling and colour command codes into documents by hand, because the word processor didn't know how to talk to the printer properly. I suppose it gave me a leg up on HTML when that eventually became popular.