ATI mach8/mach32/early mach64 Documentation?

It’s a long shot, but I’m looking for programming documentation for ATI’s mach8/mach32 and early mach64 chips (prior to 1996 or so). The earlier documents may have only existed in paper form. These used to be available from ATI but that was a long time ago. I’d be obviously happy to pay for shipping and other costs if physical items were involved.

I’m especially interested in mach8 and mach32 register references and programming guides, plus any sample code ATI may have had (I have the mach64 SDK and sample code disks, but nothing for mach8/mach32). I do have several documents describing the latter-day mach64 derivatives (Rage XL etc.) but am interested in the earlier chips, mach8/mach32 and the earlier mach64 GX/CX chips.

This entry was posted in ATi, Documentation. Bookmark the permalink.

31 Responses to ATI mach8/mach32/early mach64 Documentation?

  1. Michal Necasek says:

    I found that site a few months ago… but thanks! It’s a little too “new” (nothing about mach8/mach32 at all), though still very helpful.

  2. Michal Necasek says:

    It’s not unhelpful, but far from what I’m looking for. What I really want is the original document(s) that the brief information was distilled from.

    Also, please assume that if Google finds it, I already did πŸ™‚

  3. Mike Kerpan says:

    It’s not original documentation, but you might want to look for old copies of the XFree86 source code. It had drivers for the 8514, Mach8 and Mach32 and if those are decently commented, you might be able to find out some info.

  4. Michal Necasek says:

    It’s certainly helpful. I also have the source code to the Windows and OS/2 drivers.

    But I’d still rather read the documentation that the people who wrote the code had πŸ™‚

  5. Rauli says:

    It’s not from the makers, but I have a file called VGADOC4B.ZIP with a text file for each of the most popular VGA cards in 1995. It includes ATI.TXT with some information about the mach8, mach32 and others. The ZIP file is still available at some FTP sites.

  6. Rauli says:

    Note: The ATI.TXT file included in VGADOC4B.ZIP is far more extensive than the ATI.TXT file included in “PC Games Programmers Enciclopedia 1.0”

  7. Michal Necasek says:

    That’s the same file posted as a link in an earlier comment. My reply to that still applies πŸ™‚

    While on topic of what I’m not looking for, I should include Richard F. Ferraro’s Programmer’s Guide to the EGA, VGA, and Super VGA. I have the 3rd edition and it does cover mach8/mach32 but lacks detail. Which is not to say that it isn’t an excellent book!

  8. ampharos says:

    You should really archive this stuff and put it in the world, publicly. Eventually, these forgotten FTP archives might die.

    Archive everything.

  9. Michal Necasek says:

    I’m certainly working on the archiving part πŸ™‚ In fact I have quite a collection by now. What I unfortunately don’t have is the time to convert a ton of files into something usable.

    What I’m actually especially lacking is some sort of sensible management tool. There are thousands of files, in some cases many versions of the same document (think Intel programming documentation), some files fairly large. I doubt WordPress is up to the task, but at the same time it needs more than just a directory hierarchy.

    Suggestions welcome.

  10. ampharos says:

    A wiki? You could have them editable, for whatever reason (annotating, fixing errors, reformatting into markup), use categories, etc.

  11. dosfan says:

    How much documentation do you have and what topics are covered (i.e. DOS, OS/2, PCs, Intel x86 architecture) ?

    Please don’t do a wiki, wikis are the worst. Having a database which anyone can edit is a bad idea because it allows people who don’t know what they’re talking about to post misinformation and it also allows rampant vandalism. Wikipedia is basically a crap pile when it comes to scientific and technical information since there is no true peer review from knowledgeable people with legitimate credibility.

  12. Michal Necasek says:

    I too am skeptical of a wiki, unless it were a restricted access one (in which case it largely loses its raison d’etre). Spam/vandalism is unfortunately a problem… it only takes one or two idiots to spoil it for everyone else.

    How much do I have… I’d actually have to tally everything but I’m guessing it’d be at least 2,000 files, likely more. A few GB of data, not that big in terms of the number of bytes. But enough documents that they need to be organized somehow. The real question is how.

    Plain HTML would be perfectly adequate for describing the files but is a pain to update. On the other hand I’d like to keep things really simple and also usable in an offline form.

    The content is primarily (but not exclusively) hardware and strongly PC oriented.

  13. Rauli says:

    I would not trust any software to organise it. You and me are are interested in similar topics. Maybe not very similar, but very similar in comparison to, lets say, a lawyer or a plumber. And your classification would be very different than mine. So, what could any software do?
    Then we have the “impurity” problem: It is hard to find a single document which goes just under one category. Most cover some categories. The solution? Store it in the main category and in the others put a link, or a see_also.txt file. But sometimes the “main category” is not a single one. And if you rename or move or do something with the file you must remember to update the links or files…
    Next: The “misc” problem. If you ever allow the existence of a “misc” or “various” category, and hope that after some time you will spread its contents to the right categories and it will be empty… instead, with time it will be the most crowed category. And the most searched one: Always that you don’t find a document you will search this category. So it will become the most important one πŸ™‚
    I’ve spent an important part of my life classifying different types of information and I still haven’t found a good solution, not even a “good enough” one.

  14. Michal Necasek says:

    My rough idea is to structure the documents by manufacturer/organization (think bitsavers PDF archive). No room for ‘misc’. And probably use some sort of tagging to allow alternate “views”. That way the stale links issue does not exist.

  15. Rauli says:

    You also mentioned different versions of the same document. But what do you do when you receive the (possibly) same document version in different formats? For example, in PDF and Word. How can you make sure that it is the same? (so as to discard one). You can visually compare them only if they are short and the page boundaries are the same.

    And, about Intel docs: Sometimes you have the same document, in the same format (PDF) but different format versions (PDF 1.1 and PDF 1.3)

    The same would happen with HTML documents: The content can be the same, but not the HTML code (blank lines, different attributes order, different charset…)

    For any hashing software they are different documents. Is there a way to know that they are the same?

  16. Michal Necasek says:

    Different versions pose a problem, and especially comparing PDFs is difficult. Luckily it’s not something which comes up that often. If there are two different scans of the same documents, human intervention is probably required.

    I actually have very few HTML documents — nearly everything is PDF. Word documents almost have to be converted to PDF as they’re too hard to read otherwise (newer versions won’t even open old Word documents).

    Anyway, none of that is a serious obstacle.

  17. Yuhong Bao says:

    “(newer versions won’t even open old Word documents)”
    By default, this file block can be lifted by editing the registry or in Office 2010 or later using the UI. I agree it is a good idea to convert anyway.

  18. Rauli says:

    Conversion to a “canonical” form would make it easy to compare (possibly) different documents. Not scanned documents, of course, from which I tend to archive the largest file hoping that the image detail is better, but it’s not always the case.

    About classifying by maker, where would you put de facto standards made by agreement of an ad hoc group of different companies? (not by ISO, IEEE… etc, which could be considered like a maker) I mean El Torito Specification, or similar docs.

    Also, I would easily find old documents, which I know the maker by heart, but it would be a problem to find the recent ones, as I forget the maker soon. Does it happen just to me or is it a funny behaviour of human memory?

  19. Michal Necasek says:

    El Torito would go under Phoenix. IMO that’s the most logical place for it, even if that may not be the first place one would look. But that’s where tagging would help (BIOS, CD-ROM, etc.).

  20. Stiletto says:

    Pretty sure MAMEDEV/MESSDEV has some info on these chips if nobody’s got around to sharing anything with you yet, Michael. That said, it might be identical to the NetBSD docs mentioned earlier. I’ll take a look.

  21. Michal Necasek says:

    If they have something non-public, I’m interested πŸ™‚ (If it were indexed by Google, I’m sure I would have found it.)

  22. Stiletto says:

    Thanks for trading, Michal. Hopefully we locate that mach8 info one of these days!

  23. Michal Necasek says:

    Yeah, I’m sure someone, somewhere has those old manuals…

  24. Valery says:

    I have Rage II+c/VT4 documentation. AFAIR, I had CX manual, but can’t find it πŸ™

  25. Michal Necasek says:

    Paper form? PDF? Are you willing to share? πŸ™‚

  26. Valery says:

    It is a PDF file.
    Rage IIC and ATI-VT4 Register Reference Guide
    Technical Reference Manual
    P/N: RRG-G03600 Rev 1.01

  27. Valery says:

    E-mail me, if you are interested to grab it.
    Also, I have some other PDFs.

  28. Michal Necasek says:

    I see, recently added. Too bad I have this document already, thanks to one of the OS/2 Museum blog readers πŸ™‚

Leave a Reply

Your email address will not be published. Required fields are marked *