Talk:List of home computers by video hardware
|WikiProject Computing||(Rated List-class, Low-importance)|
Work in progress?
Is not everything on wikipedia a 'work in progress'? Let me first say I totaly agree that this would become a gigantic list, or rather a rather large table, as I intend to use a table not a list, the list is just a temporary solution. But be assured that I won't let others do the work, I intend to do it myself.
I do have the information available, I have a rather large library of older computer magazines and other material, so (most) of the information is available from a relaible source. Other detailed information can be found on the web or from the relevant pages of each system/
At the moment my main concern is how to convert the list to a table, but that is a technicality. I intend to create something like List of Intel chipsets, and I do think that it is of encyclopedic value.
Of course you might disagree, but a table (list) like this is not an unicum, there is even a category for lists like these. I simply want to compile available information in a simple to use table so thatgives an overview of the technical details of the video hardware of old home computers. The list will likely become shorter than the "list of home computers" because many variants of systems share the same basic video system. Mahjongg 17:59, 11 January 2007 (UTC)
I agree that this page should be kept. It will serve a similar purpose as List of home computers by category, which has been around since 2003, but by video hardware, not be CPU. Madlobster 20:21, 11 January 2007 (UTC)
- Aren't even paper encyclopaedias technically WIPs? The version you buy is just a snapshot of where their research was at on publishing deadline day.... :) Any of them that don't continually update and refine become history books. I daresay if this had a "publishing deadline" it would be tidied up into smart form pretty fast - OR quietly dropped from publication until it was ready for a later version. 22.214.171.124 (talk) 19:11, 24 October 2008 (UTC)
|“||this would be a gigantic list, is a work in progress, and WP:NOT a directory. Have a nice day!||”|
Systems still unassigned to a category. (work to be done)
The following list is used to keep track of systems that are still to be added. Each time a system is added from this list its entry will be removed. Mahjongg 10:43, 28 August 2007 (UTC)
Sony SMC-777 Sord M23 Sord M68 Thomson TO9 Thomson TO8D TRS-80 Model 3 TRS-80 Color Computer 3 Pentasonic PROF 80 computer (a French TRS-80 clone that used a rare video processor the EF9365) philips vg 5000 http://www.old-computers.com/museum/computer.asp?c=249&st=1
- I've shortened this list a little- please remove anything that sold originally for more than $2,000 since that class of machine was not aimed at the home computer market. For example, I took out all the "Cromemco" line since these were never intended for little Timmy to play video games or do his math drills on. --Wtshymanski (talk) 04:37, 25 December 2007 (UTC)
- That's possibly a bit arbitary. Our first home multimedia PC cost more than that, but it was definitely aimed at Home & SoHo use, including education and recreation, rather than big business. I daresay some pack-ins for the more powerful of the home systems pushed towards or even beyond that level at launch. 126.96.36.199 (talk) 19:09, 24 October 2008 (UTC)
- The Atari 8-bit family (400/800/XL/XE) machines have GTIA as the main video chip. It is responsible for generating playfields (i.e. the actual screen contents, like the screen background and pixels with different colour), players and missiles (hardware sprites), palette etc. The display can be generated by feeding the GTIA directly with data provided by the CPU. However, it is normally not done so. Normally, the GTIA is fed with data by another chip called ANTIC. This is a sort of a microprocessor, has the DMA and shares the bus with the main CPU. According to the Display List (a video-display program created and stored in the RAM by the OS, whenever screen mode change was requested) the ANTIC fetches the screen data from the memory, interprets them, and then tells the GTIA (through a separate, 3-bit bus linking them) what color it has to display at a time.
- So in reality this video hardware consists of two chips, where one (GTIA) is a video shifter, and the other (ANTIC) is a microprocessor which provides DMA and automated control of the former.
- No, it seems not many people are adding info about the systems listed above, except myself, and I must admit I have been busy with other things lately. Its just a question of time. When adding a system it takes quite a bit of research. I will add new systems piecemeal now and then. Next up is first in the row, the Commodore PET (I own one). Mahjongg (talk) 00:02, 28 February 2010 (UTC)
Well, with the Thomson TO7 done this Herculean task seems to be almost finished. If anyone has any suggestions for any home computer that I have forgotten feel free to ask (but a Wintel PC doesn't count as a home computer, as mentioned before) Mahjongg (talk) 23:06, 19 September 2010 (UTC)
Added the FM-7 other (rare (outside their home market)) systems may follow.Mahjongg (talk) 15:22, 25 December 2010 (UTC) Done KIM-1, added SOL-20.Mahjongg (talk) 16:18, 25 December 2010 (UTC) added NEC PC8001 Mahjongg (talk) 18:27, 25 December 2010 (UTC)added MUPIDMahjongg (talk) 21:24, 25 December 2010 (UTC)
- The FM-7 has been done. Main problem was to figure out whether it used programmable logic. Still not completely sure it didn't, but I guess not. Mahjongg (talk) 12:47, 10 September 2011 (UTC)
- The whole concept of a home computer is "sort of dated", as home computers were a thing of the eighties and early nineties. The picture is therefore quite representative of the genre. Mahjongg 10:50, 11 October 2007 (UTC)
IBM PC not a "home PC"?
- Because this is the article "List of home computers by video hardware" and a computer that can be used in a home is not the same as a Home computer!
- Most first IBM-PC (clone) video-cards used the MC6845 but that is in no way a criteria that decides or even is a factor in the decision whether a computer is a "home computer" and not a Personal computer. There were many other computers similar to the IBM-PC at the time, see List of early non-IBM-PC-compatible PCs. but almost none of these really counts as a "home computer" either.
- The term home computer is an old term, and does not carry the meaning you might now give it. Please read the article home computer to get an idea what makes a home computer a home computer, then you will see that the general consensus at the time when the term was in common use (the eighties) was that an IBM PC-clone was a business computer, not a home computer.
- If you look at the List of home computers, you will see that the only two computers from IBM that are regarded as home computers are the IBM PC junior (IBM's first attempt at a home computer, and much later the IBM PS/1 aka IBM Aptiva. It is possible that I will add these two to the list, but there are much more common home computers that have to be added, so it does not have a high priority for me, but If you want to add these two, then please go ahead.
- Maybe a home PC, certainly not a Home computer, which is a special term! I may stick in a simpler two cents: It's because pretty much all of the computers in this list were made to one particular, fixed design, supposed to be a cheap, almost disposable one-piece that had set capabilities. The PC was both a lot more expensive, aimed at business, and crucially was on a backplane-like expandable bus system. The video card was not set. Any PC you could list that fufills the proper specifications would be able of using pretty much any one of a list of compatible 8, then 16, then 32 bit video cards. Plus it was cloned like crazy - very few of these other machines were licensed, and when they were often little changed, sometimes not even the model number, just the manufacturer badge. You could maybe include the 8-bit card types strictly compatible with the most popular variants of the early PCs without making the list too big, but certainly once you reach the 16-bit AT era all bets are off unless you want the list to cover a small book's worth of space. It was always MEANT to be customisable, so it doesn't have a particular thing we can put in the list. Hell, you can run it without any video card at all and communicate via DOS and a teletype if you want. I ran a 486 that way via my dot matrix printer once when the monitor broke and I needed to copy some important schoolwork onto a set of floppies. 188.8.131.52 (talk) 19:05, 24 October 2008 (UTC)
ST uses video shifter
- Moved the citation below from the article to here, because it belongs here, not in the articles body.
(Actually, this is not entirely true, as the Atari ST also uses a video shifter chip as well. In fact, it is far more advanced than the CDP1861. So, the CDP1861 is really NOT the only video shifter chip to exist in a home computer.)
- It might be true that the ST also had a chip that was called the "video shifter", but that is irrelevant. in the case mentioned (CDP1861) the "video shifter" was the -only- video hardware used. The ST's video hardware contained much more than just a simple sync signal generator and shift register. If I count the ST chip as a "video shifter", Ill have to do that to halve the other video chips too. Ill try to make the point more clear in the article. Mahjongg (talk) 16:44, 11 December 2007 (UTC)
I think your statement that "the 64 used the weekest of the mentioned processors" should be removed as it is misleading. If anything, most people would agree that it is the most powerfull of the three. Especially because of the pipelining that allows it to operate at almost four times the speed of the Z-80 at an equivelent Mhz rating. —Preceding unsigned comment added by 184.108.40.206 (talk) 19:15, 20 March 2008 (UTC)
- I did not unintentionally used the prefix (arguably), as complete "wars" were fought at the time, over which processor was "the fastest", or the "best". What makes a CPU fast, is much more complicated than the simplistic take you put on it, even forgetting that indeed the 6502 was only available in 1MHZ versions, while the Z80 had a 4MHZ version. But so many other factors are more important, the general architecture, the number of registers, the flexibility of "addressing modes", the stack architecture, far too many factors to single just one out as you do. But as a matter of fact I am agnostic to the issue, My first system used a 6502, and my second a Z80, but the general understanding of the time was that the Z80 had a "more powerful and serious" architecture than the 6502, which was seen as a "cheap copy of the 6800", and that the "6809" had the best instruction-set (especially stack management) but lacked in availability of software. All that is (understandably) debatable, but that was what the general take on things was around 1985. Enough said, for every person who claims the 6502 "is the fastest", another will claim with the same amount of zeal, that the Z80 was "better". In the end, it seems that it counts more which CPU had the most and the most zealous users than any "facts". As far as I know the issue has never been resolved in any meaningful way by a neutral body with enough clout to finalise the matter, so the statement expressed here is just a "general indication of the common denominator of publications etc of the time.".Mahjongg (talk) 21:58, 20 March 2008 (UTC)
Contradiction in the article
The C-64 video chip, the VIC, is classified as "Video interface controller" in the section "The main classes of video hardware". But the C-64 below is not listed below ("The list of homecomputers, and their video capabilities") as a system with a video interface controller, but as a system with video co-processor. I think the latter is an error: VIC is not a microprocessor (it doesn't feature own instruction list, unlike the video controllers in 8-bit Atari or Amiga), it is an I/O chip, so the Commodore 64 is simply listed in the wrong table. It should be moved one table up. 220.127.116.11 (talk) 14:45, 5 June 2008 (UTC)
- The term "co-processor" in this context does not mean it is a "co-micoprocessor" but that it "processes" data by it's own means, it for example detects a sprite collision by its own means. So co-processor (in this context) doesn't mean it processes instructions from a list, just that it can handle some tasks independant from the main processor, and that it co-operates with the main microprocessor.... as far as I know only the Atari had a "display list processor", in the sense you give to co-processor... Note that the co-processor article has this to say about this... "A coprocessor may not be a general-purpose processor in its own right. Some coprocessors cannot fetch instructions from memory...." If you do not agree please discuss this issue.. If not I will revert your action in time... Mahjongg (talk) 22:47, 6 August 2008 (UTC)
- Your definition includes ZX Spectrum's ULA for example, because it "handles some tasks independent from the main processor" (= fetches screen data and processes them), unlike, for example, on ZX-81. So, it seems to me, that according to your words, virtually almost everything can be considered a video co-processor, including, of course, the VIC. But this makes little sense, so it would be more sensible to make a narrower definition of a video coprocessor, as a chip that behaves like a procesor on its own, i.e. has own instruction list, own program counter and so on. But that's not the main point here. The main point here is inconsistency in the article. The section "The main classes of video hardware" lists C-64 as having "Video interface controller" (citation: 3. Video interface controllers were a step up on the ladder, these were true VLSI chips that integrated all of the logic that was in a typical CRTC based system, plus a lot more, into a single chip. The VIC-II chip is probably the best known chip of this category). The tables, however, list the C-64 not in this class, but with systems featuring video processors. Mamurra (talk) 10:43, 7 August 2008 (UTC)
- In afterthought: if displaying sprites and doing automatic collision detection makes a video co-processor, then 8-bit Ataris have two of them: ANTIC is a processor, so it can't be classified otherwise, and GTIA does sprites and collision detections, so, following the definition above, it is a video coprocessor as well... But in reality such classification makes even less sense, so it seems to me that it is better to say that sprites/collision detection do not make a coprocessor, VIC is what its name says (Video Interface Controller), and C-64 belongs to the class of system with "Video Interface Controllers" (as the article claims otherwise). Mamurra (talk) 10:54, 7 August 2008 (UTC)
- Whats in a name... I was just trying to get some "hierarchical" order into all the different methodologies used by video hardware to create a Video signal. And -my- "definition" of a co-processor was, when I think back to why I made the distinction is that its a self-contained unit that can operate independently of the main microprocessor, and simply gets commands from it through some kind of "pipe" after which it does all the work (work meaning the manipulation of the video RAM content, so that the content of the video signal (video picture) changes, or that some events that happen in the picture is detected automatically, for example a sprite collision, or the finishing of complex manipulation of the VRAM contents, on its own. Obviously what the co-processor does should go beyond the basic, i.e. "fetches screen data and processes them", that all video display systems must do, as its a step up on a hierarchical "ladder" of other "less capable" or less integrated systems. Or as I have put it, "Video co-processor chips are Video interface controllers that can manipulate, and/or interprete and display, the contents of their own dedicated Video RAM without intervention from the main CPU". So they not only do everything a "video interface controller" does, that is doing the work of a CRT-controller (creating the timing and reading the video RAM content to create a video signal from it) and doing all that in a single chip, not a bunch of discrete chips, or with one CRT-timing generating chip (for example the 6845), and a handful of logic chip. A co-processor does/is all that, -and- does it without being completely dependant for everything it does on the processing power of the Main CPU system, it can handle some things without CPU intervention. With "things" meaning it should handle the manipulation of the video RAM on it's own as a co-operating system. Yes the ANTIC does that AND also can read and interpret "instructions", so in theory its another step up the ladder, its both a co-processor and a co-central-processing-unit (for lack of a better word).
- But you were right! I agree! You could argue that the VIC (and VIC2) chips are a bit in a grey zone, in that they did have sprite collision detection logic, but everything else was done by the main CPU directly controlling the internal registers of the chip. So however potent these chip were, You are right that they never should have ended up as a co-processor. A clear example of a true co-processor was the TMS9918 which operated all on it's own, with its own independent memory, which the main CPU could not directly manipulate, it could only give the co-processor orders to do it for him. I was a bit curious myself why they (VIC and VIC2) ended up a in the CO processor table, especially as the VIC2 (C64 video chip) is the canonical example of a Video Interface Controller and I can see that I myself indeed put the VIC chip in there, but not the VIC2, that edit was done by a user called chip8. I agree both should move from the co-processor table to the Video Interface Controller table, NOT however the Amiga, it really did have a true video co-processor! Now the first amiga is in VIC table and the newer are in the co-processor table which makes no sense. I will correct that, and the VIC20 and C64 will stay in the Video Interface Controller table where they best belong. Mahjongg (talk) 15:19, 7 August 2008 (UTC)
I disagree with your assessment of the TMS9918 as being a "true co-processor" and placing it in a different category than the VIC II. The biggest difference between these chips is that one has a separate memory bus and the other shares the CPU memory bus. I don't see how that could be connected to the question of whether something does or doesn't qualify as a co-processor. They both depend on the CPU to setup the display parameters and data and both will continue to produce the same image after the CPU stops making changes. I believe that a more important, and certainly more clear, distinction is to be made between the video hardware which only passively generates the display and those that can actually modify memory. Neither VIC II or TMS9918 ever modify memory, AFAIK. Even ANTIC, although it has a more comprehensive means of configuring display than just a few registers, does not modify memory and doesn't really do any "processing" other than calculating an address, reading data, and passing it on to GTIA at the right time. Contrast these systems with the Amiga, which has a blitter that can modify memory, as well as a "copper list" which is executed from memory and can modify the chip's own registers (including those that control the blitter, which in turn can modify the copper list!) I would put the 8-bit Ataris, and the TMS9918 systems in the same table with the C64. In the "co-processor" table should be only the Amiga, MSX2, and any other systems that have the ability to do block copies, draw primitives, or otherwise make changes "behind the CPU's back." —Preceding unsigned comment added by 18.104.22.168 (talk) 07:03, 27 January 2010 (UTC)
- I hear you, lets see what coprocessor has to say about this. I quote "A coprocessor is a computer processor used to supplement the functions of the primary processor (the CPU)". About the kind of graphic coprocessors we are taking about here, it says. "Another form of co-processor that became common during this era were the simple Video Display coprocessors, as used in the Atari 8-bit family, the Texas Instruments TI-99/4A and MSX home-computers, which were called "Video Display Controllers". So they even explicitly mention the TMS9918 (as used in the TI-99/4A and MSX1) as a coprocessor. I don't think that "modifying memory" counts as a prerequisite to be called a coprocessor, but actually the TMS9918 -does- manipulate the video memory contents under the control of the CPU, you give it a command to place some data somewhere in the Video memory (through the I/O mechanism of the Z80) and the TMS9918 accepts the command and places the data in the video memory. But as I said, manipulating memory isn't a prerequisite for a "coprocessor". It's true that the definition of coprocessor is a bit vague in this context, I tend to see it like this, if it can accept commands from the CPU, and can carry them out independently from the CPU, then that is one distinguishing hallmarks of a coprocessor. Another one is more of an architectural difference, a coprocessor normally has its own internal "Microarchitecture", (the interconnected structures that "do the work") that it does not share with the CPU. When the architecture of the video chip is heavily interwoven with that of the CPU, we tend not to consider it to be a coprocessor, but a "peripheral". Mahjongg (talk) 00:23, 28 January 2010 (UTC)
I don't quite get the video hardware definitions and the placement of some of the video chipsets: "Video co-processor chips are Video interface controllers that can manipulate, and/or interprete and display, the contents of their own dedicated Video RAM without intervention from the main CPU." A quite weird definition. If anything which interprets and displays data is a video co-processor, how comes that not every Video Interface Controller is in the Video co-pro list? Because VICs also interpret and display data. What does "manipulate" mean? Does it mean changes in RAM? Or is it just a different word for "interpret"? And why dedicated video RAM, especially when several of the chipsets in the video copro list don't have dedicated video RAM? 22.214.171.124 (talk) 18:11, 17 December 2010 (UTC)
- well, if you think you can come up with a better definition of a video-co-processor, I am curious as to what you can come up with. Mahjongg (talk) 23:28, 17 December 2010 (UTC)
Amiga text mode
The Amiga has no text mode at all (in the sense of storing and displaying an array of characters in/from video memory). The only modes it can display are planar bitmaps (140, 70 & 35 ns pixels, color depth of 1 to 8 bits), all text characters need to be rendered to bitmaps. If there are no objections I'll correct the table. -- Zac67 (talk) 14:56, 27 November 2010 (UTC)
- I'm well aware the Amiga had no hardware text rendering features (I own several models) as you can see from reference #55. If you think you can add anything relevant go well ahead. The "text sizes" mentioned are sizes supported by the GUI OS, in combination with the screen resolution the GUI uses as a default. Obviously the smaller the Font bitmap size the more letter you can cram onto the screen at a certain screen resolution. Mahjongg (talk) 16:42, 27 November 2010 (UTC)
- I agree that fast video content manipulation, or hardware tricks to emulate them, can be quite essential features, but in fact impossible to put into the article as "table entries"! There are simply too many incomparable variations on this theme. The only way to give any or all of them any mention, is to add more references in the special features sections. Go right ahead if you want to add pertinent details, but dont go overboard with it. the notes section is already overly long. Mahjongg (talk) 12:57, 10 September 2011 (UTC)
- I added a hardware acceleration column to all tables where you can add tokens for one of currently four hardware acceleration techniques:
- SC for hardware supported (soft) scrolling
- BL for blitter support
- SP for hardware sprites (not using software or blitters but real sprite hardware with z-buffering)
- DR for hardware accelerated line drawing
- I added a hardware acceleration column to all tables where you can add tokens for one of currently four hardware acceleration techniques:
I had also previously come up with a "color resolution" entry, but the hardware acceleration column also inspired me to change the "lowercase" column into the more versatile "font extras" column. As for some time I was hop[ing to somehow incorporate things like TRS-80 block graphics, and PET's pseudo graphical characters systematically into the tables, as many systems were inspired by the TRS-80 and PET to do something similar. If anybody knows of any system predating the TRS-80 and PET in using these tricks please do mention this here! Mahjongg (talk) 22:47, 16 September 2011 (UTC)
- P.S. I seemed to remember that the SOL-20 also used a few pseudo graphical characters. I looked it up and indeed the SOL used something called the "6574 character generator ROM", (made by Motorola) in which the first 32 of the 127 available characters (control codes in ASCII) were a collection of line drawing characters and such, which probably makes the SOL-20 one of the very first systems to use them.
The question is whether the 6574 was specially developed for the SOL-20 or was a pre-existing chip. Anyway, the SOL-20 has been incorporated in the discrete logic table. Mahjongg (talk) 00:57, 17 September 2011 (UTC)
- P.P.S. It seems the MCM6574 Motorola character generator ROM was developed especially for dumb video terminals in a time that the bitmaps for such fonts were still normally built using diode matrices. ROMS with more than a few dozen bytes were still rare at the time. For the bitmap of the ROM see  page 2-126 (the MCM66740 was the equivalent of the MSM6574) Mahjongg (talk) 10:57, 19 September 2011 (UTC)
While i was at it, I have also compressed the three column's detailing details of the sprite system into one column, hopefully it will make the tables a bit compacter. Mahjongg (talk) 22:36, 17 September 2011 (UTC)