These VGA Cards have a COPROCESSOR!
Вставка
- Опубліковано 11 чер 2022
- We take a look at some cards that reportedly support the IBM 8514/A standard. These cards have a hardware blitter.
We also take a look at a Paradise CGA card from the mid 80s with Plantronics COLORPLUS support.
8514 Demo Diskette:
mail.lipsia.de/~enigma/neu/ibm...
Video about Ron Belewski's 8514 Mahjong Solitaite:
• Ron Balewski - Mah Jon...
Images:
IBM VGA Graphics Card - Vlask - CC-By-SA 4.0
MCA Slots - K Reichert - CC-By-SA 3.0
Amiga 500 - Bill Bertram - CC-By-SA 2.5
creativecommons.org/licenses/...
creativecommons.org/licenses/...
creativecommons.org/licenses/...
Other images used under fair use doctrine (e.g. non-performative work or for educational content) - Наука та технологія
BGI (Borland Graphics Interface) I think is just a way to abstract away how to draw on the screen, but it probably isn't going to be really using the low level features. I also had BGI support for Turbo Pascal. It is just setting up the drawing mode and providing a library which could be used for drawing business graphics for a time when most PCs still weren't running Windows.
Yes, I think that's right.
Definitely keep the channel as a hobbyist thing. As long as you can pay the bills and put food on the table through regular work then relish the freedom!
Completely agree. The "for-passion" channels are so much better quality content, and they attract a really dedicated group. Nothing mainstream can compete with that. There's more to life than monetization.
Thank you for this great video. I have a great time learning new things about very old hardware. Have a nice holiday!
Thanks, you too!
@@PCRetroTech Alas, no holiday for me yet. Thank you anyway though.
Really find the content of your videos interesting and enjoy them (along with your jokes!), and am just glad you're going to continue in whatever capacity you enjoy.
Thanks.
Any capacitor across the power input, especially to a new\different(non-system) PC board is very important. Those capacitors also prevent interference from LEAVING the video card and polluting the other slots power-supply or motherboard chip-set. Those by-passing caps are always important.
The reason SCART had to support RGB was probably because of the SECAM color standard the French used. SECAM modulators are more complex than PAL or NTSC modulators do to the differences on how color is modulated. With RGB, you just needed some resistor ladder DAC.
Early Dendy Famiclones in Russia used to have a SECAM PPU made by UMC, but later dropped it for a PAL one because it was cheaper.
That sounds about right. Europeans love complex standards and then standards that try to combine all the previous standards and so on. SCART really looks like something designed by a committee of experts, none of whom agree on anything. 🙂
I just stumbled on your channel. Love watching the in depth technical stuff. As for topics, one I'd like to know more about (and which strikes me as right up your alley) is were there video cards (ISA or PCI) that generate an interrupt (apparently INT2) on vertical refresh? I know you can poll some registers to detect vertical refresh but I've read around the place that INT2 on vertical refresh may have been a thing. Anyway, if you could find some time to hold forth on the subject I can guarantee you at least one viewer.
Regardless, thanks for sharing your hobby and knowledge about it.
There's actually a video about this on the channel. It's entitled "8 Bit Computer Graphics Technique Almost Unknown on PC?"
That Paradise chipset was used by Commodore in their PC clones. They first used the PVC-2 that is on your card, and in later models they moved to the PVC-4, which afaik is the same, but integrated in a single chip. It's very similar to the ATi Graphics Solution, and later the Small Wonder.
Both the Paradise and ATi were often found in clones, as they were relatively cheap, and a single card supported both monochrome and colour monitors, at the flick of a DIP switch (the 16 MHz oscillator is for MDA/Hercules mode, where CGA runs off the timing signal from the ISA bus. Hercules support is also why it needs 64k of memory).
The "smooth scrolling" they mention on the packaging is probably a reference to text scrolling in CGA mode, which has snow on a real IBM card. These cards don't suffer from snow.
I think they may have used that parallel port controller as well. I think I read that somewhere.
Yes, now that you mention it... The IBM MDA card came with a printer port, and this was the common way to add a printer port to a PC in the early days. MDA clones, including the Hercules, would also offer a printer port, probably for that reason: to be a drop-in replacement. I see that this Paradise card also comes with a printer port. That's one area where it differs from the ATi Graphics Solution/Small Wonder.: it doesn't have a printer port. The Philips PC clone I have, which came with an ATi Small Wonder, has the printer port on the motherboard. So I guess there was this transitional period where some clones added certain peripherals on board, but leave the video as an expansion card.
The later Commodore machines have the PVC2 or PVC4 on board, as well as the printer port. But I wouldn't know if they used the printer port that came with the Paradise chipset, or if they left that bit off and used their own solution for the printer port.
i remember all the hard work making blitters and such working through proprietary api and ports back in the days ^^
VgaDoc (4b) from Finn Thoegersen was a goldmine of information when you were out of luck getting your hands on proprietary documentation.
The one doc I remember from back in the day is Ralf Brown's interrupt list.
Very enjoyable video. The all mighty algorithm you speak of popped this one upon the home page for me and I'm glad it did, very interesting topic and I look forward yo your future video when the other card arrives.
Thanks. Glad you enjoyed the video!
Man I love stuff like this, while PCs are quite standardized these days I've found you can still find quirky hardware and novel attempts at new things in industrial components. Lots of innovation going on in the microcontroller space. I've been looking at industrial imaging/vision stuff lately too and while it's outside of my price range I see lots of novel hardware approaches there as well.
A lot of people who are interested in this stuff are working on microcontrollers or embedded hardware devices for a living. There's a certain mindset for sure.
This video is truly fascinating!
Thanks.
Very interesting topic!
Thank you, too for the really interesting content. Even if I'm not so interested in graphic card's, I highly enjoy your video. Please continue to talk about, what is really your hobby! It makes videos so much more enjoyable.
Thanks, I've noticed that with other videos too.
The divide error is actually a problem of many Turbo/Borland Pascal programs. Essentially it means that your CPU is to fast. There are programs that can slow down your CPU to avoid that problem, and in principle you could patch the delay calibration routine that causes this problem.
Hercules graphics cards had 64K of memory, and support page flipping. But from what I've seen that wasn't widely used even in programs that supported Hercules graphics. This was exploited by dodgy manufacturers who would put faulty/marginal/reject chips in the other page to save money. It would work fine in most programs, but any that used page flipping would glitch out. The user might assume it was a problem with the program, since it worked fine with other software, not realising that it's caused by faulty graphics memory.
The dirty tricks!
I was going to college in 1988, and the art professor had an IBM PS/2 Model 60 with the 8514/A and the 8514 monitor. He would open up he office and just let me hang out and work on the computer. I was an Amiga and Atari 8-Bit guy, and this taught me a lot about what else was out there. This machine had a 4MB 3rd party memory expansion on the micro-channel bus. That couldn't have been cheap. There were serious problems with memory management on this system. Friends of the professor kept sending different memory managers. I eventually found drivers that came with the system disks that fixed the problem. Finally!, EA Deluxe Paint running in 256 colors! Also, rather than OS/2. this system was running Windows 2.1. Getting Adobe Illustrator and Aldus Page Maker playing nice was also a challenge.
Ah good old Deluxe Paint. It's worth all the effort.
Sounds like you art professor might have been into doing art electronically?
Whilst we had the best computers money could buy when I was in high school, they were BBC Micros and eventually an Acorn Archimedes, which unfortunately the students understood better than the people maintaining them. Students were barely let near them except for boring classes as a result.
@@PCRetroTech The school was almost all IBM. I think the Computer Science department had a 360. Everything else was either original 5150 PCs or PS/2. The art professor I was told did artwork for IBM executives. It's an area of Florida where the IBM top brass probably all owned homes. The professor really didn't know much about computers, but we did drag the machine into the studio to do some live drawing of the model. We also had a large plotter that was donated. It had a IEEE-488 interface. I didn't see much chance of finding a MicroChannel solution to the problem. :)
My 2nd graphics card after a trident vlb was a mach32 pci board that had truecolor I think is what really amazed me. I later sold it to get a VLB card that did.
Your cards are different revisions to the ones on VGA Museum. If you have the time, it'd be worth submitting them to be added.
Another highly enjoyable video and exciting travel updates - I know the retro community here in the UK would be very happy to have you here! Very happy for you to keep this a hobby channel whatever works best for you - it does bring a lot of freedom to cover whatever takes your fancy. I really enjoy the old obscure gfx stuff when things were still up on the air, maybe you could look at other weird and wonderful early PC card experiments: sound cards, media, etc... Also perhaps check out some of the reproduction stuff like the graphocs gremlin (maybe you could help TubeTime with his Tandy/PCjr Graphics BIOS code he was struggling with?) So much cool stuff to cover, cheers!
I'm based in London btw 😁
@@mogwaay Cool. I won't be far from there if everything works out. Thanks for the encouraging words.
And I'm sad, that he leave Germany. It's a pity, that our regulations make it hard for him to stay in Germany. Congratulation to the UK's retro community.😉😁
@@olaf8175 I'm not sure the UK is much better. I think it is just a general problem for anyone on a visa. One advantage with the UK is I speak the language better and can more easily figure out the rules.
Ooh you mentioned my favourite card of the '90s - the Tseng Labs ET4000. I had the W32 and it was so good for raw speed in games.
Yes it was a very popular card for that reason. I have a few Tseng cards.
Perhaps the 64KB RAM was used to provide the flicker-free scrolling ..
Actually, Hercules Graphics Card has 64 KB of video RAM, so the card had to provide as much in order to provide 100% compatibility with HGC.
One 720x348x1bpp HGC screen takes 32 KB, but HGC supports full 2 pages of graphics (with the ability to disable one of them if it needs to co-exist with a colour adapter, as far as I remember).
Ah, that would explain it! That explains that then.
Divide by zero usually means the CPU is too fast. Usually when made with an early Pascal build. I believe there are a few patchers that fix this
That would be surprising in this case. The author actually expected it to be run on a fast 486 which is what this is.
@@PCRetroTech Same thing happens with Jazz Jackrabbit when you try to run it on faster machine, yet game clearly suggests Pentium for highest audio quality.
Anyway usually slowing CPU with turbo button or disabling CPU caches in newer machines is enough to get such software to run without patching it.
@@Leeki85 Ok thanks. Then I'll give it a try. Maybe that's all it is.
@@PCRetroTech Some more technical details: The game does QueryPerformanceCounter() and then divides the result by a value take from QueryPerformanceFrequency(). The division is done in way which requires the result to be 32-bit. If QueryPerformanceCounter() returns a value which won't fit to 32-bits after division, the game will crash with said error (yeah it's NOT a 'classic' /0 issue)
Ah I see. Thanks for the extra detail.
Excellent video! That is very fast for BGI, I used it too back in the 90s (high school) and BGI on an ISA ET4000 was pretty slow.
It seemed fast to me too,
@@PCRetroTech I wrote a "demo" for high school computer science using BGI and it was dog slow on my 200MMX with S3 Trio V64 (2mb). Did not even try it on our 486 DX2/66 with Tseng ET4000. So I am very glad you got that performance
@@Rouxenator i wrote a rotating cube with one color faces using bgi fillpoly function it didn't look very slow but actually was) dx4-100 in my school
Great video!
Thanks!
86box has emulation now of the 8514/a , which I have tested and works well with windows and os/2 8514/a drivers. It even worked in my testing with the xfree86 8514 driver. Using this card in windows, os2 or early Linux (xfree86) would be an interesting test.
Oh that's interesting. It would be interesting to try that out. I wonder if it just emulates the 8514/a at the interrupt level or tries to emulate the register set.
@@PCRetroTech I suspect register level because Linux does not use the interrupts to access the card, and certainly doesn't use a tsr or dos driver; it uses the card at a register level. And as I mentioned, xfree86's 8514 server does work in 86box.
@@ryaxnb2 That makes sense. I guess it must be at register level, which is interesting.
@@ryaxnb2 xfree86's driver doesn't use the accelerating features of the 8515 and Mach 8. It simple falls back to VESA SVGA.
@@OpenGL4ever interesting. I have since tested the emulation with os/2 2.0, which also works. I suspect os2 does use some acceleration features.
Really cool video, love the IBM 8514/A. Now I feel the need to add support to FastDoom. Would be cool Doom running on an accelerated ISA card.
Yes it certainly would. I even thought of writing an optimised triangle routine for it that recursively broke it down into rectangles.
Hijacking this comment to say FastDoom is an awesome project, and everyone should check it out!
@@bryanpratt3933 That's the sort of thread hijacking I fully support. And I totally agree!
The Headland HT208 (a Video Seven 1024i clone) is supposed to support 8514/a but I don't think it supports it fully. I was able to get Windows 3.1 working with the 8514/a driver on a computer which had integrated Headland graphics.
Interesting. I hadn't heard that.
i physically installed an 8514/A card for an office worker who thought it would make her computer faster. She was was fairly disappointed.
Ah ha! Yes, I imagine the IBM marketing material had that kind of effect on people with way too much money. LOL.
I vaguely remember having an ATI Wonder video card with the mach32 chip. NOt sure what I did with it. LOL
1280X1024 in 1992 sounds insane
Big fan of the graphics coding videos. Just had a thought about how it it might be interesting to see a kind of 'demos - deconstructed' series. I've no idea how some of the magic was done in some of the more famous demos like Future Crew's Second Reality. I realize how difficult that might actually be, but maybe not for you? Anyway, good luck with the new gig in the UK!
Thanks. I'm not entirely sure how accepted it is to publicly disassemble demos and give away the secrets. Having said that, the source code for Second Reality is available, simply because it is so famous and Future Crew decided to release it.
The way I like to do things is to try to replicate what was done simply by observing external behaviour. A lot of demosceners privately disassemble and even share source code of course. That's entirely accepted as the norm as far as I can see. But for much of my professional life I've not had access to the source code of competing products and disassembly would be impractical and illegal, so I've developed a skill of figuring out how stuff is done without seeing the code.
I agree such a series would be highly interesting of course. I'd certainly watch it. But it's basically encouraged to only reveal how your own code works, not the code of others. It also becomes messy when people share their code with you, so you know how it is done and you are supposed to keep it a secret. This happens often.
@@PCRetroTech Absolutely and totally respect that. Thanks for everything you do and have demonstrated so far!
I do like your videos, they are informative and technical but not too much so. So I love that you've struck a great balance here. My thought about monetizing, is it so, that you can't monetize even to your home country? I heard from someone that UA-camrs can choose their channel's monetized country.
That's not possible now due to cross border data sharing with even banks being required to report income to the country of tax residence as of about a year or so ago. Those UA-camrs are going to find out the hard way that it is illegal.
@@PCRetroTech eek! I'm glad you're on top of things.
You could probably use a magic eraser to get rid of that stain. There's the cheap variant melamine foam I think... Good luck
The paradice VGA WD90C30 was actually the last one in the run to not have a blitter. The 90C31 was considered a "windows accelerator". And the performance difference on a 386 DX33 with the proper drivers in windows 3.1 is quite shocking. The 90C31 promises to be 8514 capable, but i don't know that it will do any better than anything else you tried. I think some people confuse that for simply being the resolution and color depth, without understanding it was a standard of programming.
Another card I would expect be a worthwhile try is the CL5426 and 29. I have the latter somewhere but i dont know off hand.
There's some very good documentation available for it, and I am looking forward to trying one of those out one day. I imagine it works the same as the ATI card in many respects. I don't know what other features they have but many of these were faster than actual 8514/A.
You know, since I have heard about 8514/A back in the 90s... this video gave me more than 100 times the info compared to what I have learned at the time.
Great to hear!
Sometimes old programs give divide error when running on too fast cpu. I don't know if that's the case here with mahjong, but just something that came to my mind.
It turns out that the program was expected to run on a fast 486, so I wouldn't expect that to be the problem here. I think it just didn't work on a wide range of hardware as it was never really finished off by its author.
For cleaning, a melamine sponge, aka Mr. Clean Magic Eraser will do it. Though you have to be careful to not also remove some paint. They work by being abrasive and having a higher hardness than what you're cleaning. just be gentle and it'll usually go fine.
Thanks, I'll give it a try.
@@PCRetroTech Seconding the magic eraser - slightly wet, and be gentle as commenter said. Good luck!
Did some testing with my own cards. None work with this driver.
Tested the following chipsets on a 386: ATI Mach64, ACUMOS Avga2, S3 P86C801, S3 P86C911, Cirrus Logic CL-GD5434, two different ET4000ax's
I may have to test this with a vlb mach32 I have...
Thanks for testing those out. I believe there are drivers out there for the ET4000 which will work, but certainly not the IBM one. It may be some work tracking it down though.
Something that might work for the stain (I should test it on the stain that the tantalum from an IBM CGA left in my Turbo XT case) would be a powder bleach or similar. Since those can take off the metal residue from cutlery on ceramic bowls and such. The residue that's left is likely vaporized tantalum metal.
Sounds like it could work. Thanks!
That's actually not a "smoothing" capacitor, but a bypass one. The idea is that since the power supply is a long way away, but the ICs need quickly varying amounts of current, it makes sense to store current close to them. This way for the high frequency components of the current wanted by the ICs, you have a local source, essentially bypassing the long way to the main power supply. Boards can work without them, but it's not to uncommon for them to be very glitchy without them.
Ah thanks for pointing that out.
Technically correct, but I've seen the terms used interchangeably by hobbyists. It makes sense if you consider that without bypass caps, your local rail would end up very "unsmooth" as the IC gates switch. You could even call them "filter" caps if you consider the supply rails a signal or reference from which you must remove high frequency switching noise. So perhaps a bit pedantic, considering UA-cam audiences?
@@bryanpratt3933 I'd seen them referred to as a smoothing cap before, but I'm happy to know about the correct term. It's more or less about how you hold your head when you say it, but if there is a more correct term I don't mind using it.
What about the S3 cards? I thought they were based on the 8514 as well.
Very loosely I'd say. At some point I'll take a much closer look at their blitter.
The bgi support is pretty neat. Would be fun to know if it accelerates or is just support/useful in that combo with vga in a ibm machine for hires
Edit: i think bgi has bitmap copy too of some sort
I wondered the same thing about BGI. I'll try to figure it out in the next 8514 video on the Mach 8 card.
The Mach 64 evolution of that ATI core hung around for a LONG time. It was the 2D core for the onboard Rage XL chips found in server mobos into the early 2000s. Also of note: the ATI accelerators (at least the Mach 8 and 32) are about the ONLY ISA accelerators that work on the 286 specifically because they're 8514/A clones, you just use the IBM drivers that come with Windows 3.0 and up. I'll wager OS/2 1.3 is the same way. As far as I know all the other accelerated ISA cards (the Cirrus Logic ones, S3's offerings, the ET4000/W32i, etc.) require a 386 to run the drivers.
That's interesting. I had not thought at all about the drivers requiring a 386. I'm glad to hear the Mach 8 ones work on a 286 as I certainly can't afford an ISA Mach 32.
@@PCRetroTech which one did you get? The accelerator only version (the Mach 8 proper)or the one with a standard video chip on it (Referred to as the ATI Graphics Ultra IIRC, typically has up to 1 meg of ram for the accelerator and 512K for the regular video chip). The driver issue also means that most really advanced ISA cards are kind of wasted on a 286 because you're just using the "Dumb framebuffer" capabilities of the card , which may be suboptimal. Supposedly the Graphics Ultra is one of those cards (but it handles ISA bus OC'ing well) as are some of the Cirrus Logic cards. Someone on Vogons did a fairly comprehensive comparison of high end cards on a fast 286 (0 wait states at 20mhz). I've heard that one of the better dumb framebuffer cards on the 286 is none other than the Trident 8900D with the full 32 bit memory interface. The plain non accelerated ET-4000 is still the best from what I've heard, but finding one with a good RAMDAC that doesn't cost an arm and a leg is tricky. Most of the 8900D's I've seen, including mine, have the nice and fast Trident DAC. If I ever get finished with my Xi8088 build and get back t o my 286 build I'd like to compare the 2 (the ATI card and the Trident).
@@ironhead2008 I got the one with the VGA core, so it won't need a separate VGA card to run.
@@PCRetroTech It might be interesting to check the DOS performance. Supposedly its subpar there, even though that core on its own without the accelerator onboard (I forget what they called that card , it had half a dozen names depending on whether it was OEM or sold by ATI themselves and whether the RAMDAC is good or is a potato) is actually quite a good card for DOS.
@@ironhead2008 I'll try to figure that out in the video. I'm not expecting anything to impressive. I read some reviews that were not flattering from memory. They certainly weren't aiming at taking the gaming crown.
Sort of a left field question, what set of drawers / shelves were you using to keep all those cards organised?
Just cheap plastic shelves with very cheap plastic drawers, kinda shoebox size.
@@PCRetroTech Thx! Simpler than I thought, beats my giant stacked IKEA drawer of AGP cards
Divide error rings a bell - wasn't there a problem with Turbo Pascal programs that fail on computers that are too fast?
Yes, a few people mentioned that. It's very likely the issue.
Micro channel, where the PC industry completely rebelled from IBM’s leadership!
Android smartphone is amazing good job Sir
My Schneider EuroPC supports COLORPLUS too :) but to bad no software/demos for it.
Yeah it's another dead standard in so many ways. Killed by EGA I guess. Some of those pieces of software listed on that Paradise box must support it though, one thinks.
@@PCRetroTech yeah, I already have an image viewer and some sci games that were patched, but there is no demo that uses plantronics. there is the CGA-Demo that kind of tricks with the background colors, but this is for 4,77Mhz and uses PCSPK, if you have a faster machine like my EuroPC then you have about double the performance available.
@@drzeissler Ah I see. It's a shame they didn't write something for it when they manufactured it. Probably they just got a cheap deal on the chips which happened to have support for those modes.
Oh man. I had no idea the Mach8 I got for $34 in 2020 could do that. I vaguely remember something about it supporting the IBM 5175 PGA monitor, but having never seen one come up for sale, I somewhat forgot about it.
I’m curious how that monitor fits in with these standards, if it does at all. Are you running it connected to a VGA monitor? I was probably mistaken in thinking I needed a PGA monitor for that mode to work. 🤦♂️
It's just a standard VGA monitor. In fact I'm even using an LCD in this case. The high res mode doesn't work properly there, but the low res one does.
I’m probably stating the obvious here, but 8514 hi res is 1024x768 in glorious flickery interlaced mode. LCDs often don’t support this mode at all, or correctly.
The 8514 monitor, I believe, only supports VGA and 768i. No 800x600. IIRC. I have one of those monitors but haven’t gotten round to servicing it yet, so I haven’t tried this myself.
Taking about coprocessor on video cards, are there any plans for stuff like TMS34010 or TMS34020..
Seem then a lot on magazines in late 80s and really 90s... Never in real life...
I bought one, but so far have been unable to get it to work. Their troubleshooting guide didn't help solve the issue, unfortunately. Finding time to do extended troubleshooting is proving difficult at present.
@@PCRetroTech ah, I see. Maybe that's why there's virtually no videos about them on UA-cam?
@@maniacaudiophile No I think that's just down to people mainly just showing code other people wrote, rather than their own code. That's understandable. It takes a lot of effort to write code oneself.
There was one other odd standard that IBM came up with. PCG or PGA. 1984 ish? It was STUPID expensive.. but many motherboards still referenced it when you set it to EGA/PGA/VGA mode
Yeah I wonder if I'll ever see one of those. I'd also like to see a TIGA one day!
25:13: That particular image, as nice as it is, is actually imported from Macrocom's ICON DEMO, so it's not Jim's, and it's not actually as complex as it appears at first: Much of it is just deceptively simple ASCII art shenanigans combined with that cell height register ÷ 4 hack. The same picture appears in an eponymous video on my channel, in glorious potato quality.
Have you tried MEK electronics solvent?
No I will look out for that.
@PCRetroTech be careful, it's more toxic than acetone, use glove or long swab, well ventilation recommended.
@PCRetroTech: Something that has frustrated me for years is for the many graphics cards released with accelerators few of them were used for DOS.
During the Windows graphics accelerator years which ran concurrently with the DOS games era many accelerators were reviewed with benchmark figures for Windows GDI drawing. I haven't seen any examples of people finding a way to use those Windows accelerators for DOS games. Like today I am not even sure if the manufacturers released the spec.
It would be great to see what performance those Windows GDI accelerators would get when being driven directly from DOS. It may also have opened up a new avenue for DOS games to take advantage of graphics acceleration features only seen in Windows. Even the Tseng ET4000 cards had acceleration.
What chance is there of making drivers for those accelerators which could be used from DOS?
One problem is that DOS is 16 bit, without protected mode unless you use an extender. It's pretty nontrivial to do that sort of thing by hand. Windows as an OS was delayed many years before its initial release and DirectX didn't appear until much later still. It's just not easy to build up all the driver layers, make it compatible with the zoo of PC hardware that's out there, document it, test it, support it and make it flexible enough that people will want to use it.
What about Patreon or PayPal donations? That's different than channel monetization.
No, the issue is with the type of work (UA-cam presenter) not the money. That's why vintage programming might be a possibility (though it requires registering a business and I am not sure if HMG allows it or not).
Speaking of video cards with coprocessors I have a card with a "TIGA" chip, aka Texas Instruments TMS34020 (there's also the older TMS34010) which is probably one of the very first actual "GPU" made over a decade before Nvidia created that name. This is a complete CPU that was made with graphics in mind. On PC it got used on a few video cards for CAD, but from what I've seen, it was more popular in the arcade scene. It was used to power games like S.T.U.N. Runner or Mortal Kombat.
As for ideas for future videos, your Paradise card with 64K of ram got me thinking. It turns out that my ATI small wonder also supports plantronics mode and also have 64K of memory, so that makes me wonder if something is possible with that extra ram on the paradise, maybe it will on the ATI as well ? What if there's some "hidden" standard built over the legacy of the plantronics mode ? (Kind of like EGA where most cards had 256K of ram even though it was originally made with 64K). I need to check which other cards are compatible with plantronics and how much ram to they embed. If they're all 64K, maybe this could lead to something quite interesting (CGA-class hardware with scrolling would be amazing !!)
Hopefully I one day encounter a board with TIGA support.
I will certainly be looking into that 64kb, but someone mentioned that it was probably for scrolling in Hercules.
By the way, was it you that asked me about XT cases. I seem to have ended up with a spare, albeit without a badge and with a small crack which I haven't looked into repairing. If you are still looking for one of those feel free to email me.
@@PCRetroTech oh that's awesome !
Now I'm moving from one place to another and I'm currently trying to make more room than using it so I'd rather wait a few months if possible (if not, well I'll see what I can do).
I happen to have a TIGA card, however it might be broken :(
@@Dxceor2486 I'm moving as well, but so long as it is out of here before about the start of September it should be ok. It was part of a lot that wasn't free, but we can discuss those details via email when you contact me. Just let me know when you have moved.
@@PCRetroTech email sent
Btw Fastoom have plantronics support as well other modes. Might be usefull for speed testing as it has integrated fps counter for demos. Tested on it plenty of Hercules, CGA, EGA and some VGA cards....But im not sure how much is cpu limited as i did testing on my Athlon XP2200+ which isnt exactly cpu meant to be paired with ;-)
Btw some Tiga or even older cad accelerator card (were some today unknown prececessors) would be nice to review. Sadly i'm not aware of any TIGA compactible game for PC (but were used for example in Mortal Kombat arcade cabinets).
Oh, I didn't know about that. Thanks for letting me know. I think one of my viewers did actually mention FastDoom, but I think for something else.
I really wish I had a TIGA card, and even more a monitor which worked with it. But I definitely don't have access to one at the moment.
@@PCRetroTech Think i have 2 working TIGA cards and both are ok with VGA monitor, they even works on mine LCD screen without issues. But both are late 90s models...so monitor might not be issue (with right card).
@@vgamuseum Ah interesting! I didn't know the cards were still around so late. Another one I'd love to play around with one day is a TARGA graphics card. Actually, maybe that's what I was thinking of when you said TIGA.
@@PCRetroTech Well i didnt know that targa has its own standard....only one i have is late 90s video editing card....
Tiga cards sometimes appear on market, main issue are drivers as each company had their own, so card without them is uselles. Always check before buy, that you have source for drivers....
@@vgamuseum Oh cool. I haven't seen one come up here where I live but perhaps they come up in the US. And thanks for the tip with the drivers.
Yeah my co-processor (wife) is away atm so I might reboot...
@7:16 Don't get yourself fooled by Wikipedia. I've seen ET4000AX cards running 8514 software, so there is some truth to it, but the actual truth is that the ET4000AX is a simple non-accelerated framebuffer chip. Tseng Labs provided a pure software implementation of the interface provided by HDILOAD, called RIXAI4 (for 16 colors) and RIXAI8 (for 256 colors). Thus, I consider it a stretch to call the ET4000 chip "8514/A compatible".
@8:00 The original 8514/A is 16-bit MCA, which will fit the IBM PS/2 Model 50 perfectly, which is also a 286-based machine.
@22:22 The 16.00 MHz oscillator on the "Hi-Res CGA" card is for the mono modes. The original IBM MDA used 16.257MHz, and all the clones, including the original Hercules card, used 16.000MHz. I don't think there is any kind of microprocessor on that card, though.
@23:05 It's not a BIOS chip. It's an EPROM chip, but it is the "character generator" aka "font ROM". The IBM compatible PCs include the BIOS required for CGA and MDA modes in the mainboard BIOS, so these cards typically don't have a BIOS.
@23:20 While you would not expect 64KB on CGA (or even the colorplus), 64KB is the standard memory amount on any Hercules compatible card (which this card is, if it is set to mono).
That's what I concluded too. I've only found three register level clones of 8514/A so far: ATI, WD and C&T. That's going to be an expensive video to make, LOL.
Already on ic system
Adapter only caed need run on power
Divide error. Try slowing down you computer. Just a random guess, that it is a timing loop, and CPU is too fast.
This card came out much later than the machine I'm using here, so I don't think timing would be the issue, but it's worth a try I guess.
Pharmaceutical petroleum for that stain on the case.
Thanks!
1.
24:48: Am I seeing that right, that there seems to be a slight kink in the rows, i.e. they're not arrayed entirely straight? Is that something on the monitor that causes that?
2.
11:15: Massive respect for how you REALLY did your homework - and ALL the homework.
3.
Having recently watched various unrelated retro console videos, the CGA light pen connector now makes me wonder if there were ever any light guns available for CGA PCs, and if there were any games that supported those? I've never heard of either, not for CGA PC/XTs, but I'm wondering...
1. Yes, there is a kink in the rows. It's not the monitor but just a minor bug in my code which I didn't find time to fix.
2. Thanks.
3. Unfortunately I don't know much about light pens. I've never seen one in real life.
@@PCRetroTech (re 3.) I'm mildly curious as to whether anyone ever had and used any one of these with CGA. Some googling just now led me to a comment on reenigne's blog, where Trixter said he'd been looking for a light pen for two decades. That was 13 years ago...
To my understanding, a light pen is fundamentally similar to a light gun, only the latter has a lens and is intended to be used at some distance to the monitor, while the former was supposed to touch the glass.
I think the closest I ever came is, I saw some DIY light pen build instructions once, but I... no wait, actually those instructions are also linked _ibidem,_ in another comment. Google Basics of Light Pen Operation. So maybe there never was a vendor? Maybe everyone made their own? Still, the other question is, does compatible software yet survive? I'll have a look-see next, whether there are any UA-cam videos on light pens...
@@ropersonline I did think about making one even if I have to write my own software. But they aren't very accurate reportedly, so a bit of a gimmick. You might be able to press some buttons with one, but I'm not so sure about actually drawing with it.
@@PCRetroTech PS: Apparently my above understanding was wrong at least for some light guns. According to Wikipedia and an older 8-Bit Guy video on the subject, light pens work as expected, but light guns use a variety of methods, like looking for a simple flashing square as in the case of the NES model.
@@ropersonline Ah I see. That makes sense.
Since I do not have anymore my huge boxes with old hardware, I would like - if you can and have time - to make a list of ISA16 VGA adapters that work on an ISA8 slot motherboard (XT,286,386 or 486 used to have an 8 bit ISA slot). I needed some times to use an VGA card in old systems so I used 16 bit ISA cards on 8 bit ISA slot and some have worked. Also tried with some VLB boards on plain ISA motherboards and some have worked, with some missing/reduced functionality (I remember a dual IDE VLB board lost one IDE channel and DMA capabilities when plugged on a motherboard without VLB).
And about that dual video card feature, can you show how it worked and what it was used for? I used a Hercules mono clone with VGA on my 486 when I programmed on Pascal; it sent the debugger on the mono card. Could use CGA color with a VGA adapter on the same time or only Hercules mono? Were there other applications/programs/games other that CAD/Lotus that took advantage of the dual video card setup? And how about Windows? How it handled two video cards? I know that Windows 3.1 did not supported Hercules mono (or was 3.11?) without some tweak. Also I remember that the BIOS screen looked different when using mono/CGA or EGA/VGA video. It would be interesting to show how a WinBIOS look with various video adapters.
I think different memory addresses were used with monochrome and CGA so I think that's how multiple cards worked, but I didn't try it. Unfortunately my monochrome monitor is still broken so I can used the mono setting until I get it repaired. The signal is not compatible with a CGA monitor.
You really need to get some static bags. I got 100 off ebay for like 7 dollars, to store these cards in. I used to store them in boxes just tossed in, and messed up a few. Its always sad when you been a bad steward and killed stuff due to carelessness.
All my valuable and rare cards are stored in antistatic bags. But the cards also aren't going to die just sitting on the shelf. It's how they are handled that matters.
SCART is a french/german invention.
20:59 Try wd40 on a paper napkin and rub it.
Thanks.
Nobody cares.
Get a Real Job.
Thanks youtube for the: Dont recommend channel option.
I wrote a comment on one of your other videos about IBM 8514A regarding another 8514A matter - very long comment and IMHO full of very useful information and it disappeared, i.e. it was deleted - I don't know why and if you deleted it or UA-cam deleted it, but in either case it is very bad, because now I am extremely reluctant to waste my time in writing any more comments. Anyway, I am going to give it a second and last chance with this comment. So, I want to tell you my opinion (based on the error you got) why it fails with your AGX014 card - I bet that card has only 512 Kilobytes of memory. So, let me elaborate, 8514A memory is organized in 4 or 8 Planes each of them 1024 bits size, i.e. 512 Kilobytes or 1024 Kilobytes (1 Megabyte) of memory. IBM sold 512 Kilobyte cards only in the US, HDILOAD you have is from IBM diskette for the German market, i.e. it's for 8 Planes, there are different version of HDILOAD - one for 4 Plane and one for 8 Plane. So, that particular IBM HDI load will work only with cards with 1 Megabyte of memory. So, it doesn't mean/prove AGX014 is not 8514A compatible. I hope my comment is useful.
I didn't see the comment in question, but UA-cam does delete all comments automatically which contain links. It used to show these to me before deleting them, but now I don't even see them. If it didn't contain links then probably it was due to their broken AI getting confused, e.g. if you put the executable name on a filename and it thought this was a domain name.
As for the AGX014, I don't recall which one this is now, but if it was the XGA one then it is definitely not compatible with the 8514/A driver since it is not register compatible with 8514/A. XGA was deliberately not compatible at the register level on the part of IBM.
I will revisit that card at some point and try to figure out exactly what the story is. I will certainly test your hypothesis. I do have access to the US version of HDILOAD. I don't have time to check now, but according to VGAMuseum AGX015 is a higher clocked AGX014 and they say the AGX015 is based on something called RISC Engine. It can do blits and fills and the like, so my guess is it supported the AI through a card specific driver only. But I'll try to confirm when I get back to that card.
Thanks for making two attempts at your comment. I very much appreciate your time, and it was an interesting comment.
I found a moment to check the datasheet and the AGX014 is the 16 bit version of the AGX015 and both are XGA implementations, not 8514/A. So there is almost certainly a driver somewhere that provides the adapter interface, but it certainly wouldn't work with IBM's driver or with any register level 8514/A programming.
@@PCRetroTech AGX014 card is the one around 10:30 seconds in your video. It error out with code "80", after you run HDIDEMO, even HDILOAD seems to load. So, there is no US version of HDIload, because in the US both 512KB and 1M cards were sold. So, in the US you get 2 versions of HDIload - one for 4Plane (512KB) and one for 8Plane(1M), in the rest of the world IBM distributed HDILoad for 8Plane only, because outside the US they were selling only 1M cards. So, my other comment was again about HDIload, but this time not IBM made ones, but ATI made, as well as other made. It's very long to write again and I didn't keep a copy of the old comment. However, ATI has several HDILoad versions - most of them for 1M cards and the other for 2M cards (because all their cards were 1M or with option to upgrade to 2M). Interesting, Using ATI version of HDIload for 1M on 2M card also fails - no matter that 2M is more than 1M. So, basically, HDIload version always needs to match the memory on the card. Additionally, there are several HDIloads that are Emulation: Western Digital made such for their Paradise cards and it's used also on MG104 and MG108 Matrox cards (it seems those WD Paradise and Matrox MG104/8 cards are compatible) and there is another HDIload emulation for Tseng 4000 cards. Unfortunately, without be able to freely communicate, i.e. without links it will take days to track them all down on the Internet, but they are available. So, you can have quite a lot 8514A AI to test, compare bugs and performance, etc. IBM has 3 versions 1.00, 1.01, 1.02, ATi - I have at least 3 versions that survived over the years (1.10, 2.1, 2.3), then WD/Matrox HDIload (1 version ) and then those for Tseng 4000 cards (if I recall there are 2 versions - 1.00 and 1.11 that are still live on deep corners of the Internet). That means totally, there are 9-10 versions of HDIload.
@@PCRetroTech IBM XGA Adapter Interface is: (8514A AI Base Function + XGA-1 AI Fucntion Set + XGA-2 AI Fucntion Set). So, each XGA AI includes 8514A AI as well. There is XGAAIDOS.SYS for DOS from IBM that serves the same as HDILOAD TSR for 8514A card. So, if those are really XGA cards, then they should work with XGAAIDOS.SYS. I guess without links the best would be to download the whole 10GB archives of the old IBM FTP or tell me your email.
@@maturaseven7790 That sounds very believable as numerous card manufacturers had versions of it. I didn't know about the different IBM versions though.
Of course these only support 8514/A at the adapter interface level. At the register level the cards are completely different. Having said that, IBM also provided the adapter interface for XGA cards and it is quite feasible that some version of that works with this card. But I don't know yet whether that was also called HDILOAD.
When I go back to the AGX014 card I'll certainly look out for versions that are designed for this card. I have a better idea what I'm looking for now and I'll definitely take into account your suggestions. Thanks.
My eaurne opinion of this video is makked by vuhshuions, and a poe-cee-bea.