• Please review our updated Terms and Rules here

MediaVision Pro Graphics 1280p

creepingnet

Veteran Member
Joined
Feb 25, 2005
Messages
1,103
Location
Reno, NV
Sigh, what a journey it has been trying to find the proper Windows 3.1x drivers for this card so I can run it in my 486 DX4. So I got this card out of an old 486 DX2-66 that was used for Computer Aided Design back in the 90's that was in the skip where I used to work a few years ago. It's a 4MB VLB Graphics Card with a Cirrus Logic CL GD-1501 chip on it, and a TI chip of some sort that goes up to 1280X1024 screen resolution.

I've hunted high and low for the drivers for almost 3 years now, using Google, Yahoo, Webcrawler, Metacrawler, spending hours in the archives at cd.textfiles.com, all looking for two files....1280win.exe, and 1280util.exe, which are supposed to make the 2 disk set with the drivers and the utilities for the card on it. Every site that says they have them (windrivers.com, ftp.pipeline.com.au, and a hand full of other places) almost always have broken links or a broken website that does not work.

So I'm wondering if anyone here happens to have the disks or know of a place where I can find the disk images of this card. I'd hate to give up on it as it's a pretty nice graphics card for it's time. Media vision is out of business, their FTP is long gone, and I've gone so far as to look for it's smaller sister card's drivers to use - the 1024 model, and even that is like finding a needle in a haysta

so my question is
Does anyone have the disks, images of the disks, or the files? Or know where I can get em'? Or does anyone know if using the Cirrus Logic drivers or TI drivers would work? And which chip would be the chip to use them?
 
Are you sure it's a "1501" -- that's not a CL-GD number I've ever seen. Pretty much all Cirrus Logic video chips are 54xx.
http://en.wikipedia.org/wiki/Cirrus_Logic#Graphics_chipsets

Maybe if you posted up a hi-res pic we could help you ID it better... Trying to find a copy of cirrus' vlchip.zip for you, since that would ID the card properly... I would GUESS at it being a 5426 or higher 542x...
 
I had one a long time ago, so my memory might not be the best...
I think the cirrus logic chip is a red herring. The board used it for vga compatibility only. For the high-res and accelerated modes, you need the driver for its own video chip.
 
Yep, the Cirrus chip is only to provide basic (and apparently quite slow) VGA compatibility in DOS. Windows uses the accelerated chip, which is apparently some sort of Mediavision proprietary thing called the MV451... it could be a re-brand of something else, but information about it is pretty scarce so it's hard to say for sure. The TI chip is probably the DAC.
 
Sorry for bumping an old thread, but could you please post a picture of your ProGraphics 1280?
Apparently there are no pictures on the net, and I'd eagerly have a look - would be cool to compare it with the 1024 version (that was designed by the 3dfx team before they left Media Vision).
 
Here are some marketing flyers I kept for the 1280. It was originally designed to be the graphics chip companion to our 386/RISC workstation hybrid. Really just a dumb frame buffer with a few latches/shifters to aid in fills and moves. The ProGraphics 1024 was architected (by yours truly) to be a Windows accelerator before I left for NVIDIA. Unfortunately I don't have any of the drivers for the 1280 - that was contracted out to a third party for development.

Pellucid.jpgPellucid 1.jpg
Pellucid 2.jpgPellucid 3.jpg
Pellucid 4.jpgPellucid 5.jpg
 
Last edited:
Wow! Thanks for the info!
Could you please shed some more light upon both the PG1024 and PG1280? Did the 3dfx team participate in PG1280 as well? If I understand you correctly, PG1280 is inferior to the PG1024, isn't it? Do you happen to have any software for PG1024? Any other info would be more than appreciated! :)
 
So did that fabled genlock addon for the PG1024 ever existed, even in proto form?
Never heard of it. But then I left just as silicon came back and before the MediaVision marketing machine got ahold of it (thankfully). I'm sure they had all sorts of follow-on products, but any genlock capability would have required a bit of external hardware.
 
Never heard of it. But then I left just as silicon came back and before the MediaVision marketing machine got ahold of it (thankfully). I'm sure they had all sorts of follow-on products, but any genlock capability would have required a bit of external hardware.
I figure there was probably something on the drawing board since all the PG1024 boards sold were populated with composite and svideo in/out connectors, so it was probably more than just some marketing department wet dream.
 
I figure there was probably something on the drawing board since all the PG1024 boards sold were populated with composite and svideo in/out connectors, so it was probably more than just some marketing department wet dream.

Hmmm. Mine wasn't - VGA out only. Maybe it was a decision made later in production.
 
ISA PG1024. Unfortunately it didn't work the last time I tried it a few years ago. Bit rot, I suppose.
IMG_2449.jpg

Bonus question. Name this card:

IMG_2448.jpg
 
Wow! Thanks for the info!
Could you please shed some more light upon both the PG1024 and PG1280? Did the 3dfx team participate in PG1280 as well? If I understand you correctly, PG1280 is inferior to the PG1024, isn't it? Do you happen to have any software for PG1024? Any other info would be more than appreciated! :)

Pellucid was a combination of engineers from Silicon Graphics and Compaq - we met from the failed ACE consortium project. We started out designing a MIPS based NT workstation and selling the remaining stock of SGI IrisVision boards. Our workstation needed a better graphics option to compete with the workstations of the day - VLB was just coming to market and the PC solutions were a little underwhelming. So we put together a simple frame buffer controller that connected directly to the MIPS bus.

When it became clear we weren't going to have a viable market for MIPS NT workstations, we spun a VLB board incorporating the frame buffer controller to court our suitor - MediaVision. There is a sordid story involving the Indian Mafia somewhere in there. So the ProGraphics 1280 was born (it could do 1280x1024 or 1024x768 ). It has some latches and shifters to accelerate fills and screen-to-screen moves. I had spec-ed out the functionality one weekend but forgot to take in to account the differences in the 1024x768 configuration so the screen-to-screen moves had to be done completely with the CPU in that resolution. Sloooowww.

Anyway, MediaVision wanted the fastest Windows accelerator from us so we immediately started work on it. We didn't have a Windows driver for the PG1280 so we contracted that out. It wasn't going to be a fast accelerator anyway, so there wasn't much point spending much effort on it. The PG1024 was designed to accelerate the windows GDI driver interface without compromise. I wrote the display driver from scratch, but since I didn't have hardware yet, I used our IrisVision product as the development platform with some firmware additions to make it a viable Windows platform. It was slow, but allowed me to get the driver hashed out and the interface for the PG1024 solidified. Some cool features of the PG1024: it didn't have a CPU accessible frame buffer. This allows the hardware designer (Sellers) to optimize the backend memory interface without regard to CPU interaction. The register set was designed so that there wasn't a penalty for 24 bpp over 16 bpp, and equivalent to 8 bpp in most cases - even over ISA. RGB data was transmitted as 16 bits of blue and green data, with the red data in the address bits (an array of 256, 16 bit values). That reduced the data traffic by 1/3 on the ISA bus. The VLB/PCI version didn't require such shenanigans, so you could just send over raw RGB data. There was also a front end CLUT so 1, 4 and 8 bpp packed data could be sent over the bus and expanded in hardware to the 16 or 24 bpp frame buffer. Most image data back then was 1, 4, or 8 bpp so this helped a lot. The hardware did all the alignment so the driver just had to set the destination registers and shove data to the chip as fast as it could. And that sucker was fast.

But it quickly became clear the MediaVision was involved in some underhanded activities so I made the decision to leave for NVIDIA, taking most of the Compaq team with me and the SGI engineers started 3dfx. That left the PG1024 somewhat untuned but it still managed to be the fastest Windows accelerator for something like 3 weeks. Such was the heyday of 2D accelerators.

I think I still have some intermediate version of the GDI driver somewhere, and the retail floppy disk that came in the box.
 
cool back story! I wasn't expecting to see PG1024 in ISA only; all the ones I've came across were VLB boards. Didnt know there was provision for PCI front end either. Nice to hear about this. Not sure how the video works out without CPU access to the frame buffer. Would be interesting to see the hardware architecture doc on the chip. Pity I chuck the board some years back; one day it just stopped working out of the blue.
 
@resman - that card: IRISVISION. Nice to see the less-common ISA version. I've had half of one for more than a decade; still hoping eventually to turn up the GE board it's missing. I have several complete MCA versions though.
 
cool back story! I wasn't expecting to see PG1024 in ISA only; all the ones I've came across were VLB boards. Didnt know there was provision for PCI front end either. Nice to hear about this. Not sure how the video works out without CPU access to the frame buffer. Would be interesting to see the hardware architecture doc on the chip. Pity I chuck the board some years back; one day it just stopped working out of the blue.

Well, PCI was more of a promise than an implementation in early '93, so it was never built, even as a proto, as far as I know. As far as a directly accessible frame buffer, that would be necessary for Windows 95 and DirectX, or DOS games, but for a strictly Windows 3 accelerator, we could get away with it. Everything went through the GDI driver.

@resman - that card: IRISVISION. Nice to see the less-common ISA version. I've had half of one for more than a decade; still hoping eventually to turn up the GE board it's missing. I have several complete MCA versions though.

I haven't had mine plugged in since I wrote the Windows 3 driver for it. No idea if it works. But it makes for a great conversation piece on my wall of PC video cards. I still have the Compaq 486/33L that was originally used to port the IrisGL library over, but none of the source code or drivers. That just didn't seem important at the time.

EDIT: I found the IrisVision AutoCAD driver install disk and was able to back up the contents if anyone is interested.


I did find the retail box that my PG1024 came in. MediaVision marketing was all about plastering stickers over the cardboard to extol the virtues of the product. So cheesy.
 

Attachments

  • IMG_2453.jpg
    IMG_2453.jpg
    55.2 KB · Views: 2
  • IMG_2455.jpg
    IMG_2455.jpg
    88.8 KB · Views: 2
Last edited:
I have just found my card, I have the original manual and floppies fro os2. 2.0 svga quickstart windows driver install and Kais Power tools for windows discs.. all unused if you are still in need...
 
Ok I have had a pro graphics 1024 from new and never used it ( I rarely use PCs, just Macs). Ive dug it out and yes the three floppies are there. One is part number 557-0247-01 ProGraphics 1024 quickstart Instalation and windows drivers disk version 1.12 1 of 1; the second isProGraphics 1024 OS/2 2.0 SVGA Drivers Version 1.1 disck 1 of 1 part no 557-0248-01; the third is Kais Power Tools Windows Version Special edition part no 557-02762-01 (It also has a manual with it)P1150789.jpgP1150790.jpgP1150788.jpgP1150789.jpg
 
Back
Top