Talk:Amiga 1200

Quote from article: "The A1200 also offered improved audio abilities which allowed for higher sampling rate for sound playback."

The A1200 uses the standard 8364 Paula chip from the original chipset. How exactly in technical terms does it offer higher sampling rates in the A1200?

Audio hardware functionality is closely related to video hardware timing, and as a result it is possible to double the maximum sampling rate on ECS/AGA machines by choosing a suitable video mode. - From Original_Amiga_chipset#Audio_features_in_general --Anss123 13:17, 29 June 2006 (UTC)

Then I think perhaps the wording and tech specs should be modified to point this out, or maybe link to the relevant info? Just to avoid any confusion.

Probably, yes, agreed. Someone? Anyone? Hey! Don't look at me that way!--Anss123 10:06, 1 September 2006 (UTC)

"The Amiga's custom chips cost more to produce than the commodity chips utilized in PCs, making the A1200 more expensive, relative to PCs, than earlier Amiga models"... I disagree with this - like for like spec PCs cost far more than £399 (in the UK at least) they were more like £700 from memory... true this became the case eventually - but it was the R&D which was expensive not the manfacture (AAA and Hombre never made it to production for example). Just my 2p. —Preceding unsigned comment added by 90.202.37.26 (talk) 21:48, 1 February 2008 (UTC)

Perhaps. I do not recall what source I used when I wrote that (think it was Dave Haynie), but perhaps I meant to say that it was more expensive relative to the PCs contra the A600. IOW, the PC got faster/cheaper while the Amiga didn't. Also, keep in mind that you needed a monitor and perhaps a hard drive to make that £399 A1200 "usable". A cheap PC generally had that included.--Anss123 (talk) 22:13, 1 February 2008 (UTC)

I found my A1200 perfetly usable hooked up to a TV, nor did it ever have a hard drive in it. I'm not just talking about using it to play games on either. --Pobice (talk) 23:12, 22 December 2008 (UTC)

True, but computers were still expensive items at the time. It was all a matter of how much you were willing to spend, what you wanted out of it, and how poor a cost:benefit ratio you were willing to put up with. If you were a cheapskate, after all, you could still pick up late-model Sinclair Spectrums/C64s and Atari 2600s from certain retailers. PCs just made more sense. Even a £400 one - bought as a plain box-plus-keyboard/mouse, therefore competing direct with the A1200's market - you'd probably get a more capable and certainly more expandable machine. At £800 or £1600, it would be completely dominated even if you bought a similar value of upgrades along with.

I'm not entirely sure about the PCs of late 1992, but put in context of my mid-late 1994 one...

Amiga: 14Mhz, rough 386DX (or higher-speed 286)-equivalent CPU. PC: 66Mhz 486DX2 (equivalent to a 68030/040 of slightly lower mhz)

Amiga: 14Mhz?? 32-bit?? main bus. Little in the way of true DMA or bus mastering, or easy/sophisticated expansion capability. PC: 33Mhz 32-bit main bus, with its own controllers making it often fully independent of the CPU - CPU, memory, disk controller and expansion ports all capable of taking control, and in limited cases talking to each other simultaneously on different segments. Memory on its own segment (33Mhz/32bit), disk controller and video on PCI (one embedded, the other in one of four slots) at 33Mhz/32 bit, legacy ports and sound card on ISA (embedded/one of three slots) at 8Mhz/16 bit.

Amiga: 2Mb of slowish RAM, probably functionally similar to 30-pin SIMMs. PC: 8Mb of 60ns RAM in two 72-pin SIMMs.

Amiga: AGA chipset; a *usable* max of 256 colours at a *usable* max of 720x576 resolution (if you could stand the flickering or had a scan doubler), or 640x480 / 800x600 in ?? colours. PC: usable max of 16 million colours at 640x480 (non interlaced), 65k at 800x600 (non interlaced), and 256 at either 1024x768 (non-interlaced, non-flicker) or 1280x1024 (interlaced). Similar 2D acceleration features to Amiga when a compatible driver used, but CPU/bus able to chuck enough pixels around unaccelerated to easily make up the difference anyway.

(NB I'm discounting the HAM resolutions as they were unsuitable for anything but static images in practice, and required jumping through complicated coding hoops to effectively display any movement. Truecolour mode on the PC - at either 15, 16 or 24 bits - was fast and simple enough to use as an everyday resolution in Windows, e.g. for photo editing, though 640x480 in 16 bit was more "comfortable" (flicker-free refresh rate, slightly quicker full-screen redraws))

Amiga: Whatever monitor you have around to plug into it, and usually the limiter on your top resolution. PC: 15-inch high frequency multiscan monitor, spec in excess of what the video card could drive.

Amiga: Double-density floppy disk drive as main storage; option for single HDD at 3.3Mb/s. PC: High-density floppy drive (with 2Mbit/s controller and space/socketing for a second internal one), 540Mb hard disk, and double-speed CDROM. Mass storage drives at 16.7Mb/s with some modicum of DMA ability, and opportunity to add two more using built-in connectors.

Amiga: Paula audio... 8 bit stereo at a little under 30kHz, typically. Limited if any recording ability. No synthesis ability to speak of. PC: Soundblaster 16 compatible (with soundblaster pro MIDI chip) - 16 bit stereo playback at upto 50kHz, thus including CD and DAT/DVD standard. Able to record direct-to-disk at this quality. OPL2 FM synthesizer chip for reasonable quality background music with very little CPU, bus, or RAM/disk overheads.

Amiga: Fixed, roughly A500+ size case with limited internal expansion potential and a horrendously weak PSU. PC: desktop box case with room for two 5.25" and three 3.5" drives, a handful of full-height expansion cards, a 200 watt PSU (never managed to burn it out), separate keyboard, and strength enough to place the monitor on top.

Admittedly the PC cost almost four times more than the Amiga, but you got decidedly more than four times the computer for it, and the real-world price allowing for inflation was somewhat lower than an Amiga 1000 or Atari ST had been at their launch. One was based on fairly fresh technology (or, I suppose, would have been cutting-edge for a home user in late 92/early 93), the other on stuff that was stale even at the time of launch. Someone with an eye for making a worthwhile purchase and buying a machine that would last long-term, rather than being an interesting toy for a couple of years would go for the PC... even if their budget could only initially stretch to a 486DX/33 with a half-meg (instead of 2mb) graphics card, smaller disk, 14" lower-frequency monitor and single-speed CDROM... all of those things could at least be upgraded quite easily and affordably, one at a time, at a later date. Along with the operating system (which was a fair chunk of the price), given that it was entirely disk based rather than partly ROM.

We upgraded to the 486 from an ST; the Falcon, despite being slightly the A1200's superior technologically and just as affordable, looked like a similarly unwise choice. Nowhere near as powerful a machine in context as the ST had been in 1985, or even in 89 when we got it (hence it remaining at least vaguely relevant and useful in general-purpose computing for another five years)... The A1200 must have seemed like an even greater disappointment and less value-for-money. It's like, you're buying a new car to replace your slightly worn Fiat 500. You could have a Ferrari for £100,000... or a Fiat Panda for £25,000. The Ferrari starts to look like a better use of your funds. Particularly when, if you ignored the shiny adverts for the F40 and perused the full catalogue, you could get an almost-as-good 355 for £50,000...

Not only would it probably have incurred huge additional outlay to upgrade the Amiga (or Falcon) to a similar level in some of those areas (monitor, CPU, memory, external CD/HDD), it COULDN'T be persuaded to compete on many of them (internal disk, power supply, graphics, audio), at least not without significant modification, rehousing, etc.

(Incidentally, that 486 survived the full term of it's original "bought for schoolwork" remit in various guises, despite being eventually superceded by a Cyrix and then games-monstering AMD K6-2 for the main family machine, and was eventually given, with a printer and some mild upgrades but no monitor, to a friend to do his university assignments with as a parting gift when we both finished our final year... it was a good use of four times the price of an A1200, which by that point already looked like a dusty relic) 193.63.174.211 (talk) 18:06, 9 May 2012 (UTC)

This seems rather focused on raw numbers and not actual usability. Keep in mind in 1992 you would be running Windows 3.0/3.1. Vesa Local Bus would still be shiny new, the 486DX2 still not available from system integrators, and PCI still literally on the drawing board. SB16 would have cost about three quarters as much as an A1200 as it was just released.

Around 1992, I had a 386DX-33. It had an ATI Charger 512k ISA VGA card, 800x600 VGA monitor, ISA IDE controller, Windows 3.1, and a 40 meg hard drive. Oh and a Weitek math co-processor, but that was entirely useless so we'll avoid talking about it. Total cost around 3000-3500 CDN by that point (mostly from the base system in 1990). By comparison an A1200 new was around 650 CDN.

The ISA video card could only really do 640x480 progressive scan and lacked ANY hardware acceleration. Larger fonts, bloated GUI design (Windows 3.x had those obnoxious "Windows Aero-basic"-style outlines around each window) basically made this the same as a TV-res 640x200 (724x239 on my 1084S) on the 1200. 256 color graphics are just as crap as running an 8 or 16 color Workbench. Also, the fact that everything had to run on the same display with the same graphics mode limited it further. I later upgraded to a Diamond Speedstar 24X for the PC, which added hardware acceleration, but video card driver instability kept it out of 24-bit mode entirely, and Windows 3.x could not handle 16-bit color intelligently at all.

The "32-bit" chips were often running in 16-bit mode, or some hacked-together quasi-32-bit mode (386enh) under DOS, BIOS, and Windows 3.x. Large pools of memory were rendered useless as Windows 3.x had little tiny 64k stacks for the User, Kernel, and GDI libraries, and limited by bizarre Win 3.x addressing requirements. FAR pointers, anyone? Windows was bloated even back then as well, so your processor, which was much more likely to be a 386DX or 486SX, spends all of it's extra, less efficient cycles plowing through Windows bloat (like, locking a region of memory so it didn't escape..) than doing useful work. You could get a great deal more work done in DOS. As a notable comparison, a fast-ram equipped A1200 with a 1-byte UART could cause the 386DX to drop characters in a Windows 3.x terminal program even though the 386DX had a 16550AF with a 16-byte FIFO buffer while transferring data via a null modem cable @38400.

The drive controllers were not so great either. They were typically PIO and a great deal slower, due to living on the ISA bus (VLB was still shiny new and required a 486, and PCI had not been invented), and lack of DMA in the PC system. A drive was lucky to get 700 k/sec. Also, the faster speed of the PC floppy controller was entirely negated by the fact that Windows 3.x (and 95+ to some extent) had to rely on BIOS/DOS code for floppy disk access, essentially shutting everything else down multitasking wise. Does it really matter if the FDC is twice as fast if you can't format a floppy disk and edit a document at the same time? CD-ROMs were rare, and there was almost no software for them.

Sound on the PC was particularly miserable at that time. There was almost no software support for the SB Pro, let alone the SB16. Most of your sound is going to be DOS apps with Adlib FM synthesis or original Soundblaster support. Single channel. 8-bit. 22khz. Nasty single-tasking DOS or Win3.x apps. Granted that an A1200 is likely going to be limited to OCS sound levels, having working stereo sound for everything was a lot more useful than the maybe-adlib, maybe-sb PC experience. 4 channel stereo at 28khz if you like numbers so much. Synthesis is tinny, low fidelity AM-quality/c64-era crap until you get into advanced wavetable stuff many, many years later. And then it's immediately replaced by mpeg audio compression.

With just a fast ram module, I could run my C compiler over a large project, rapidly switch between a terminal program connected to a shell account for IRC chatting/FTP access and a text editor, while listening to digital PCM mod files. Fast forward to 1994 with an 030-50 accel and 16 megs of fast RAM and I'm compiling projects four times as big, running a dedicated IRC client, early web browser, formatting a floppy disk, and still listening to digital audio. The Pentium 90 with Windows 3.1 and 16 megs of RAM sitting next to the 1200 is powered off as almost entirely useless.

Also what's this about DMA? A PC has like three-and-a-half almost entirely unused DMA channels. The Amiga design has about twenty or so, depending on the model. The PC design has ALWAYS been centered around processor-does-everything/interrupts-are-god design, with only the occasional half-hearted foray away from that.

You can't really duplicate early 90s era Amiga experience without going to the Windows NT4+ line on the PC side. I used an A1200 as a primary machine up until around 1998, it was very practical and it's lack of expandability was not as crippling as some people seem to think. Renegrade (talk) 17:25, 20 December 2012 (UTC)

68020 four times faster than 68000

The 68020 Amigas ran at 14MHz, while the 68000 Amigas ran at 7MHz. Is the statement accurate that the A1200 is four times as fast as the A500? --Jonathan Drain (talk) 21:00, 23 August 2008 (UTC)

The 68020 had a 32-bit buss, while the old 68000 was saddled with a 16-bit buss. Also, the 68020 has faster instruction timings. IOW 2x speedup from not having to multiplex memory reads, 2x speed up from a clock speed boost, and 2x speed up from better instruction timings. Add some inefficiencies into the equation and 4 times faster seems like an accurate enough figure. It would of course vary from program to program, with some being faster still and others slower.--Anss123 (talk) 10:37, 24 August 2008 (UTC)

Also don't forget that the 68020 has an instruction cache, whereas the 68000 does not. It's definitely a large upgrade from the 68000. Renegrade (talk) 10:27, 5 February 2011 (UTC)

Without better source I would say it is original research. Xorxos (talk) 10:09, 6 February 2011 (UTC)

Are you really going to be so pedantic as to make one of us provide your sorry ass with a link to the 68020 CPU wiki article and/or an online spec sheet? It ran at twice the speed, had twice the bus width, an onboard cache, and a few other generational improvements (extra instructions, better pipelining etc) that improved per-clock processing efficiency. It's as much of a speed bump from the 68000 as a 10Mhz 80286 would have been from a 5Mhz 8088, if not more so (or more precisely a 16Mhz 386DX vs an 8Mhz 286). 193.63.174.211 (talk) 17:23, 9 May 2012 (UTC)

Does http://lowendmac.com/tech/chips.shtml qualify as reference? instructions per second doesn't, but the same speed comparison isn't disputed there. You can also look at RC5-72 comparison database and OGR comparison db, but those are based on highly specialized code, so mileage may vary. Zac67 (talk) 18:21, 9 May 2012 (UTC)

68EC020-14 runs 4 times faster than 68000-7 only with 32-bit FAST-RAM, plain 1200 has only 32-bit CHIP witch maks it only 2-2.5 timers faster than 500. — Preceding unsigned comment added by 94.173.29.213 (talk) 02:06, 1 November 2012 (UTC)

Er, 1/ gross memory speed will only affect operations that are heavily memory constrained (with these earlier generation chips, this isn't so much the case; single-clock operations didn't become common until at least the '030 and moreover the '040 era), so long as it isn't SUPER slow (like e.g. with the 68008 or i8088) and so doesn't hold everything up REALLY badly, 2/ the bus still runs at twice the speed, and has twice the bit width, so that means 4x as much being transferred per second... Did you mean that an unexpanded A1200 runs at 2 - 2.5x the speed of an A500 which has a decent amount of fastRAM on board? Because apparently adding it to the A1200 itself can mean a general doubling of processing speed thanks to a vast reduction in bus contention between the CPU and the graphics hardware. 193.63.174.211 (talk) 14:58, 6 June 2014 (UTC)

Does the A1200 IDE controller support PIO Mode 2?

I wonder if the IDE controller in the A1200 (or 600 or 4000) actually does support PIO Mode 2, as claimed. http://hardware.amiga.hu/mod/a1200.html, http://www.thecryptmag.com/Online/16/PowerFlyer.html and http://www.amiga-hardware.com/showhardware.cgi?HARDID=1462 seem to indicate that Gayle is only capable of Mode 0. Also, the A1200 user docs describe it as an "IDE" interface, the first ATA standard not being released until 1994 (when PIO modes 1 and 2 were added, according to this). Reports I've read on forums indicate a maximum transfer rate of 2-2.5 MB/s, which is less than the 3.3 MB/s maximum for PIO Mode 0. -- Screwtop (talk) 15:07, 21 August 2009 (UTC)

It only supports PIO Mode 0. I dont know why someone claims mode 2 was supported. Xorxos (talk) 10:04, 6 February 2011 (UTC)

Seconded. PIO modes 1+ were introduced with the ATA-1 standard in 1994. -- Zac67 (talk) 19:59, 11 August 2011 (UTC)

bus width / speed, video capabilities

There are some things here which aren't clear, and/or I have argument with :)

1. Is the internal bus 16 or 32 bit? Other cheap computers using 680x0 CPUs (Mac LC, Atari Falcon) hobbled them with a 16 bit bus (presumably with some custom glue circuitry, as I can't find any mention of the chips being made in an "SX" type design). Did Commodore go the same way with the A1200, or does its low MHz rating belie comparitively good memory/chipset performance thanks to a wider bus?

2. Is the bus speed locked to the CPU speed / vice versa? It was the usual way in non-PC machines, but we may as well be certain of it.

3. Probably need to arrange the "resolutions" section a bit better, given the Amiga's typical graphical flexibility especially in terms of overscan... unless the A1200 knocked it back a bit. Maybe something along the lines of "TV/RGB monitor: upto 360/720/1440 horizontal and 240/480 (NTSC) or 288/576 (PAL) vertical resolution depending on timing/interlace mode and overscan setting, defaulting to 320/640/1280 by 200/400 (or 256/512)", though I know that's a bit cumbersome.

(I can ramp my ECS A600 up to 720x576 without losing more than a bare few pixels to overscan on an LCD TV - the menu bar and the controls for a maximised window still being usable - and it will at least attempt to display 1440x576 despite it exceeding the capability of the standard-def cable and video (de/en)coders and the WXGA panel... it might even be usable with a component video connection and the aspect set to 16:9... Built-in deinterlacer, ysee ;).

4. Were there any colour depth restrictions on the higher resolutions, such as they were with ECS? (That was 32 colour max from a 4096 colour palette in the 320-360 and 640-720 width modes, even in interlace... but only 4 colours from 64 in the 1280-1440 width mode, and the specialist 640x480 non-interlace mode. Applying similar maths to AGA suggests it might have been knocked back to 16-from-4096 instead of 256-from-16M in superhighres RGB and VGA productivity (and presumably in interlaced SVGA/XGA), unless the registers had become the limiting factor instead of memory/DAC bandwidth)

5. Was 800x600 REALLY interlaced? It would be the first time I've heard tell of interlaced SVGA. XGA (1024x768), yes, but not SVGA. What rate would it even have run at? (I've forced a video card to 768x576 at 50Hz before, for better compatibility with the "800x600" mode of a cheap TV converter, but that was STILL progressive). Low-end bare-minimum-spec-meeting SVGA/XGA monitors usually ran at an eye-scorching 56Hz progressive instead (just enough horizontal frequency to also service XGA at 43Hz interlace, or VGA at 70Hz with some tweaking; if you wanted to at least get a slightly less painful 60Hz, it needed a "flicker free" model that could do VGA at 72~75Hz). Possibly it was made to have some kind of not-entirely-standard compatibility with vanilla VGA CRTs? (480 lines at 60Hz progressive roughly equals 600 lines at 48Hz interlace, not accounting for front/back porch and sync width, etc... I think it's something I might have once attempted when briefly saddled with a plain-jane VGA monitor, to use on a PC with a decent video card and a tweaking program, but it never worked well enough to give a usable picture)

6. Other than HAM8, or setting the palette to 256-greyscale (or a 64/125/216-colour spectrum), was there ANY kind of "direct colour" or higher-colour mode possible? 193.63.174.211 (talk) 18:36, 9 May 2012 (UTC)

Claiming Amiga 1200 can run 800x600 resolution is bending the truth. Neither Commodore or the operating system shipped with Amiga 1200 had support for this resolution and using 800x600 resolution requires using special multisync monitor. The specifications in this article should only mention resolutions officially supported by Commodore and their OS. Xorxos (talk) 06:06, 10 May 2012 (UTC)

1. It's a 32 bit bus. The '020+ can happily run on a narrower bus if speed isn't an issue. 32 bit CPUs have no problems accessing 16 bit chip ressources or Zorro II hardware (or even 8 bit CIAs). The actual bus size to be used at any time is signalled by the chipset.

2. Locking the CPU clock to bus clock simplifies the design but isn't required in general.

4. Graphics bandwidth quadrupled in AGA, so it can run superhires with the maximum 8-bit depth (planar or HAM8).

5. It's rather easy: the (ECS/AGA) chipset provides three resolutions: lores (140 ns pixels), hires (70 ns) and superhires (35 ns pixels ~= 28.6 MHz) and everything must use one of these. When designing a monitor driver/video mode you can divide these 28.6 million pixels/sec into lines and frames as you like, depending on the monitor. 800x600 plus blanking intervals would require at least (my rough guess) 576,000 pixel times ~= 20 ms. So you could run 800x600 with either 50 Hz non-interlaced or 100 Hz interlaced (very possibly eliminating sprite DMA with the timing). Super72 uses 78 Hz vertical, 25.8 kHz horizontal interlaced.

6. There's copper graphics (zero bitplanes, background color modified by copper during frame scan) that's rather low res but doesn't restrict your color space - but that's an 'inofficial' mode that wasn't intended. And there's of course sliced HAM8 which probably isn't worth the trouble. Zac67 (talk) 09:53, 17 May 2012 (UTC)

Hahaha ok... I didn't actually mean to start a flamewar, and I did disclaim that I was working from the standpoint of a later, rather more expensive machine. I tried to keep the rose tinted glasses off ... and whilst I respect your side of the coin here, I think they might have got a bit wedged on the nose :D ... I will admit that, in the absence of any information in the actual article (which I clicked on Talk today to request) regarding it, I assumed a 16-bit bus, the same as the Mac which used that CPU, and as the Falcon, so my apologies for that...

I have since come into a rather more comparable machine - a 486 SX25 that was being thrown out from my workplace, having been bought at a moderate cost (it doesn't exactly strike one as a top-end model, but it certainly wouldn't have been base spec either) in, as far as I can tell, 1992. This would have been a direct competitor for yer A1200 at the time. For the time being I'll ignore the upgrades I've done to it, and mention them later.

Processor - 80486 SX, 25Mhz. I'd estimate just as a ballpark figure that it's about twice the raw speed of an '020 at 14Mhz, running code compiled from as-similar-as-possible sources. No maths coprocessor, but does run on a 32-bit frontside bus at CPU speed. Unfortunately, the only thing this seems to be relevant for is the memory (as everything else is on 16-bit low Mhz ISA, though I was able to jack it up to either 10 or 12.5Mhz just using the BIOS, can't remember which), but it does at least give good scores there. A 50Mhz part - using, incredibly, a 50Mhz FSB - would have been an option, but was presumably considered extravagant.

RAM - 4MB, 70ns Fast-page, using half of a bank of 8 SIMM slots. Seems quick enough for the system, wasn't actually able to find BIOS settings that drove it so fast it glitched. Plenty of room for all but the most demanding software (i.e. a couple decidedly post-1992 games I installed).

HDD - 160MB, IDE, 3.5". Not the fastest, think it got about a meg per second, partly hobbled I presume by the ISA-based controlled, but it did the job. Expandable to about 520MB, but I got everything I wanted to (mostly classic 80s/90s DOS games) on there without deleting any of the existing data (10~15 years worth of fitness machine data logging and random documents, plus OS and "serious" programs) and still had 10MB free.

Graphics - some Cirrus Logic SVGA ISA card, half-meg. Will do 1024x768 non-interlaced in 16 colours, and 800x600 (or 640x480) in 256 at up to 75Hz. No hi or true colour capability in windows, but as the options exist, greyed out in the settings menu, I assume it's a matter of memory (640x480x65k needs 600KB...) rather than raw colour capability. Maybe if they'd been clever, 640x400x65k could have been used. Maybe it can actually do it at lower resolutions in the few games and DOS apps that offer sub-VGA-rez hi/true colour. Who knows, haven't the software so haven't tried. The acceleration isn't the best, but it's only seemed a bit "lumpy" when running Windows and its apps at top whack (ie SVGA in 256 colours) rather than outright sluggish. It's plenty usable. DOS games, on the whole, love it, though wierdly Screamer hates it, recording only a 2MB transfer speed. I suspect that's possibly because it hammers the CPU so hard with floating point 3D sums that it hardly even sends stuff to the video card very often.

Sound. Internal squeaker. We will speak no more of it.

Regular floppy drive.

Would have come with a separate 102-key keyboard, 2-button mouse and a 14" SVGA monitor, but they were all scrapped before the PC itself was shoved in a cupboard. I replaced them with generic spares and first my HDTV, then an SVGA projector.

Software - Win 3.1, DOS 5, MS Office 4, a scad of default accessories and games, and a plethora of bespoke sports science stuff. Onto which I've piled loads of old DOS VGA and EGA games. All of which run surprisingly well. In timed tests, after tweaking the BIOS up from what seemed to be failsafe defaults - and making no other upgrades - I could go from power off to typing a document in Word in about two minutes flat. I daresay the Amiga would beat that, but few modern PCs would, and it's not a long time to wait around for your computer to become usable on the whole.

The stars of the show in this case are Tempest 2000 and Descent. The former runs pretty much like it would have done on the Jaguar. The latter is stunningly smooth, I wouldn't have expected it of a 25Mhz SX. It's a full 6-axis 3D texture-mapped shooter, if you don't remember it...

Anyway, since then I've chucked in another 4MB, a SB16 sound card, and a CDROM drive, and really it'd be a perfectly good basic computer for someone who hates the internet and digital photos/videos, and is only interested in making Office documents rather than receiving and opening modern ones from other people. The two changes I'd make to it as a classic gaming rig would be to swap the CPU for a DX-50 (or a 5V DX2/50, would make little odds), and maybe add a larger HDD, 420MB or so. Otherwise, it's good for almost anything the 90s could throw at it, including Win95 and the internet as it was in the dialup era, but short of 3Dfx-requiring games...

I have to go now, so I'll just ask - how's THAT sound compared to your Amiga, now? :) 193.63.174.211 (talk) 17:23, 12 February 2013 (UTC)

You're right, this is getting a bit out of hand, so I'm going to try to be as brief as possible. I'd love to geek out about Descent and Wing Commander II and Dune II and all of that, but I'm sure some editor would drop by with WP:Not A Chat Board and WP:Fun Is Not Allowed, etc, and delete this. ;)

I would disagree with two minutes boot being usable. My current desktop boots into LibreOffice Writer in 28s total (including typing in a password, which is probably padding that out by 3-5s..a bit more if I typo it instead of typing it). A1200s with IDE hard drives and fast ram boards can boot in about 15s, those with 030-50s can boot in 7s.

ALso I'd not call un-accelerated windows graphics usable unless you're patient and like to watch windows peel onto the screen slowly. Pre-PCI (and some PCI/VLB boards) are limited to writing 16-bit values ('WORD' in x86 assembly/windows parlance) to the video memory through the ISA bus via a tiny 64k aperture in base RAM at either a000:0000 or b800:0000 (one is text, the other is graphics). This doesn't hurt TOO incredibly much in 320x200 or 320x240/8bit as those both fit into that region, but anything larger has to use bank swapping. An accelerated machine, on the other hand, only has to receive simple directions via registers (such as http://www.atarimagazines.com/compute/issue151/110_Diamond_SpeedStar_24.php or the A1200), and a BitBLT engine does the heavy lifting while the CPU gets to go on it's merry way doing other, useful work. If you're going to do some retro computing, I do recommend that Diamond SS24X board. A1200s of course don't have to worry about bank swapping as the ChipRAM is all directly mapped (umm 00000000 to 001FFFFF I do believe) without any bank swapping. The planar architecture holds it back a bit, but then again, it holds VGA back as well (standard VGA uses planar graphics for any 16-color mode). I'm breezing past things like the multi-screen desktop with hardware scrolling support which greatly boost productivity on the A1200's side, but I did promise to try to be brief.

BTW, In DOS, large portions of Win3.x, and almost all BIOS functions, the system is running in Real Mode, limiting access to 32-bit functionality. The data bus will still operate in 32-bit mode, but all instructions will be moving to and from 16-bit registers, presenting an unnecessary slowdown that doesn't show up in the raw paper stats.

Also, regardless of what CPU mode you're in, the ISA bus is a big albatross on the back of the PC; your two megs/sec speed is actually very good. An A1200 has everything on the local bus, unless you have an accel board, in which case it's on an external, 14mhz*32bit bus (vs 8mhz*16bit, and I really don't recommend overclocking ISA busses unless your hard drive controller is certified for 10+mhz operation. Dropping a sector would not do good things for your OS or data)

Finally, some of those components may be post '92 upgrades. I found a history document on the computing industry that featured the initial release of the Compaq DeskPro 486DL/50 (it was a 486DX-50), costing $11,000-and-change, and it only had a 120meg drive. I don't see how a discount 486 could have something better than a top of the line DeskPro, so I'm suspecting that the hard drive, at least, is post-92. Most video cards in that era, by the way, usually had barely-better-than-VGA-grade RAMDACs. It's entirely possible that that machine was actually bought heavily discounted in '94 or so, or had some components replaced with discount ones around that time.

All that being said, however, I would still love to play with a machine like that. Programming on metal like that is quite fun, as the limitations turn from frustrations into challenges then. On the other hand, for usability, if I were condemned to use some older PC forever, I'd do everything I could to hold out for at least a PCI-based Pentium, or I'd be forced to avoid any and all PC GUI OSes.

Quick notes - 'XGA' is a horrible way of stating resolution (as is 'SVGA' for any specific resolution). The video card manufacturers looked at XGA cards, shrugged, yawned, and said "pass" (even though it's significantly better than most cards that bear the name 'SVGA') and the *GA notation goes quickly out of hand (WXGA+ and FWVGA and WHUXGA, oh my!). A3000s, A4000s, and A1200s all have 32-bit chip ram (although only AGA units have 32-bit display chips/no other ECS machines have 32-bit chip). I don't believe there's any "SX" style 68020s, as the 68ec020 IS the "SX"-priced chip of that line and has 32-bit data bus. That 70ns ram is effectively 14.286mhz. It's also 8bit memory(which is why it has four). My attempt at brevity reduced the size of this document by at least 50%, but yet still failed to be anywhere near 'brief' in an absolute sense. My apologies. :C

Renegrade (talk) 02:39, 28 April 2013 (UTC)

Well, OK, I might have been confused on the DMA front, and the comparison was maybe a little unfair (it was, however, designed to demonstrate just how badly the final generation of Amigas ended up being kerbstomped by the PCs that were coming along at the time, especially when the non-early-adopters started doing the maths) but I do have some disagreement with various other points here. I'm in the middle of chucking together a somewhat fairer comparison, with an older (1992), rescued 486SX I now have as a retrogaming rig... and it rather seems like THAT PC *still* offered capabilities along similar lines to the A4000 (rather than the 1200), yet may have been *cheaper*... WITH an included monitor & etc. Anyway, I'm not done yet, I've just got a notepad file slowly filling up with random notes and stuff, it really needs reorganising and editing and trimming-down to avoid another 10-screen-high nuclear data explosion, so I'll see about coming back to copy in the final thing sometime later. For now, though, that is the impression I'm getting.

Incidentally, just how many times did you ever format a disc whilst continuing to work on something else anyway? That really can't have been a terribly common occurrence. And AFAIK it worked rather better under 3.1 than it did in later Windows (and, naturally, DOS), for whatever reason. The PC drives were rather faster, though, it only took about a minute to format, you didn't even have time to go and make a cup of tea. The drive on the A600 I've since picked up also (the mechanism and logic I believe are identical to that of the A1200) seems dreadfully, ball-achingly slow, even for reading data (and indeed, even vs my old ST), let alone formatting...

EDIT: Wait, what? I thought I was responding to a section much further up. Looks like I might actually have done that updated comparison here instead. Presumably in my sleep, many months ago? WTF? Anyway, might as well double check it, I suppose...

Incidentally, FWIW, 35ns pixels versus SVGA works out as follows: a little over 900 pixel clocks per line (the highest rate VGA runs at, factfans - and it defaults to 720 pixels wide in that mode, but can be reprogrammed to show 800 on the limits of overscan / a typical analogue monitor's width adjustment), meaning a 31,746kHz line rate; modes with 600 active lines typically have 624 or 625 total (taking a leaf out of PAL's book), which takes us to 50.79Hz. The generally listed maximum line rate for the chipset is usually given as 31.44kHz (whether or not that might be a typo) - adjusting for such with the same number of lines (actually the easier option, circuitry-wise) gives us 50.3Hz, or we can have 628 total lines with a more exacting 50.05Hz vertical refresh, a slightly more comfortable 909 clocks per line, and a 31.431kHz line frequency (which could arguably be rounded up to 31.44 if you were a marketeer). 50Hz on an SVGA monitor, yum. Flicker city. 193.63.174.211 (talk) 15:06, 6 June 2014 (UTC)

External links modified

Hello fellow Wikipedians,

I have just added archive links to 2 external links on Amiga 1200. Please take a moment to review my edit. If necessary, add {{cbignore}} after the link to keep me from modifying it. Alternatively, you can add {{nobots|deny=InternetArchiveBot}} to keep me off the page altogether. I made the following changes:

• Attempted to fix sourcing for http://www.pegasus3d.com/total_share.html
• Added archive https://web.archive.org/20040330104120/http://www.softhut.com:80/cgi-bin/test/Web_store/web_store.cgi?page=catalog/hardware/amiga/a1200.html to http://www.softhut.com/cgi-bin/test/Web_store/web_store.cgi?page=catalog/hardware/amiga/a1200.html

When you have finished reviewing my changes, please set the checked parameter below to true to let others know.

Y An editor has reviewed this edit and fixed any errors that were found.

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—^{cyberbot II}_{Talk to my owner:Online} 03:59, 24 February 2016 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified 2 external links on Amiga 1200. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Added {{dead link}} tag to http://www.bombjack.org/commodore/amiga/amiga-commodore/A1200_Users_Guide_%28UK%29.pdf
• Added archive https://web.archive.org/web/20140523024513/http://www.jschoenfeld.com/products/delfina/Delfina_manual.pdf to http://www.jschoenfeld.com/products/delfina/Delfina_manual.pdf
• Corrected formatting/usage for http://www.softhut.com/cgi-bin/test/Web_store/web_store.cgi?page=catalog%2Fhardware%2Famiga%2Fa1200.html

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 00:16, 4 July 2017 (UTC)