Found the R3D/100 and more!

[EOCF] Tim

Newcomer
Finally I've found the cards, via finding the old Real3D crew.

There is a whole story behind the cards, but to make it short, R3D/100 never came to the market with only engineering samples, and demo cards produced , initially it was aimed at the mid/high end range of PC/Workstation, but didn't catch on because it was a 3D only part, to fix that, work started on the Lightning/110, which included an iXMicro Twin Turbo 2D add on PCB.

Oh and did you know the R3D/100 was initially going to have a 3rd chip on there? A complete T&L chip, due to the immense cost it was dropped. It would have been 5 years before the nVidia 256!

Unfortunately, drivers have been long lost, and no one knows where they are, so no testing, even if the boards would be fully functional (which they most likely are.)

The reason why many of you have never seen these boards is because when Lockheed-Martin sold the remains of Real3D to Intel, they were told by Intel to dispose of every single piece of hardware, documentation etc. so it wouldn't fall into the wrong hands. A huge dumpster was brought in the main hall, and was filled with everything they had been working on for years. The Lightning was close to production at that point.

One or two took some out of the dumpster to remember how much effort and time they had spend on the projects.

Apart from that, I've got a Cobra prototype which would have been released in 1999, not much known about it, I'm trying to get in contact with the engineer who worked on it. The project was cancelled, but I'm not sure if that was because of the nearing demise of Real3D or because the project didn't work out as planned.

I received some Marketing Material as well (4 caps, 5 keyrings, and poster). Poster shows the Pro-1000 image generator, and some demo's/games that were programmed for the R3D/100.

The second R3D/100 is missing two caps, repairs are scheduled for that board. Apart from that the second board has got a big scratch near the first chip, which seemingly has damaged a trace or two quite badly.

Interestingly, the R3D's got a handwritten serial number 3 and 25 on the back, so they might be from the first run, I'm trying to find out more on that.


Enjoy!!


R3D/100



#2

 
Last edited by a moderator:
I remember first reading of the joining of Real3D with Intel and being very excited
On one hand you had a company that had been doing military simulators for years, stuff way in advance of anything seen on the pc. and then you had intel a company that made cpu's which were way more complex than a gfx chip. They had thousands of staff, great facilities their own fabs and a yearly rnd budget that was more than the total worth of all the gfx chip makers put together - In my mind how could the product not wipe the floor with the opposition

And then it arrived, talk about underacheiving :(
 
Man you rock! Excellent stuff... I'm a collector too and I love to see this kind of stuff reborn from the dust.. :)
 
WOW!
Sweet cards!
From the codes on chips the silicon was baked end of 1996! Great year in which I still had an Amiga 1200.
BTW guys do you know of any PC GFX card operating on planes like AMIGA chipset did?
 
Wow, I finally saw the card. All these years I had only seen the logo. It was rumoured long ago to be a Sega Saturn add-on for porting Model 3 games.

I have seen the R3D/1000 Pro from Sega Model 3 board. But it's nice to finally see it's little brother.
 
Whoa, extremely nice update Tim! I've always wanted to know more about the R3D/100.


This is cool
199503real3dr3d100postezd1.jpg
 
I remember first reading of the joining of Real3D with Intel and being very excited
On one hand you had a company that had been doing military simulators for years, stuff way in advance of anything seen on the pc. and then you had intel a company that made cpu's which were way more complex than a gfx chip. They had thousands of staff, great facilities their own fabs and a yearly rnd budget that was more than the total worth of all the gfx chip makers put together - In my mind how could the product not wipe the floor with the opposition

And then it arrived, talk about underacheiving :(


But you're talking about the low-end Real3D & Intel designed i740 / Auburn chip used in the Lockheed Real3D StarFighter cards, not the R3D/100 chipset/board.
 
Last edited by a moderator:
[EOCF] Tim;1272686 said:

Nah, I think he means the memyry layout of the pixels, where all the bits for a pixel wasn't in one piece as it generally is today. This mode was called "chunky pixels" on the Amiga, and all the kickass graphics cards we re of that type.

As for the other mode, planar pixels. One bit for all the pixels were in a contigious are in memory, then an area with the next bit. This allowed som flexibility of the amount of bits per pixel, ranging from one to eight.



The memoty layout of the different formats is ass follows (Assuming 8 bits per pixel).
PnBm: Pixel n, Bit n.
Chunky:
[P0B0, P0B1, P0B2, ... P0B7, PNB0, ... PNB7]
Planar:
[P0B0, P1B0, P2B0, ... PNB0, P0B1, ... PNB7]

The format had one rellay great disadvantage, to change the value of one pixel, you would have to read all the bytes in the planes belonging to that pixel, change the bits in those bytes corresponding to that pixel and write all the bytes back again, for 8 bit images that is eight read-modify-write operations.
 
yes (image stolen from Swayee)

Yea, I'm not sure how much Intel was involved in the ultimate design, in my opinion Real3D shouldn't have collaborated with Intel. I mean Intels GFX cards or IGP's were always terrible, maybe their influence in the project wasn't helpful.

Saying that, I don't know too much about the i470 and the why how where when.

Thanks for that picture btw, I've never seen a Starfighter box in real life though.
 
Nah, I think he means the memyry layout of the pixels, where all the bits for a pixel wasn't in one piece as it generally is today. This mode was called "chunky pixels" on the Amiga, and all the kickass graphics cards we re of that type.

As for the other mode, planar pixels. One bit for all the pixels were in a contigious are in memory, then an area with the next bit. This allowed som flexibility of the amount of bits per pixel, ranging from one to eight.



The memoty layout of the different formats is ass follows (Assuming 8 bits per pixel).
PnBm: Pixel n, Bit n.
Chunky:
[P0B0, P0B1, P0B2, ... P0B7, PNB0, ... PNB7]
Planar:
[P0B0, P1B0, P2B0, ... PNB0, P0B1, ... PNB7]

The format had one rellay great disadvantage, to change the value of one pixel, you would have to read all the bytes in the planes belonging to that pixel, change the bits in those bytes corresponding to that pixel and write all the bytes back again, for 8 bit images that is eight read-modify-write operations.


Yes, correct!
An advantage was really fast scrolling using blitter and copper.

BTW I had CyberVisionPPC card for my Amiga1200! OpenGL acceleration was working nicely. That was fastest graphics card developed for this computer :devilish:.
 
Incredibly nice find! Sad to hear about the lost drivers, I would love to see what these babies can do. Thanks for sharing pics, nonetheless :)
 
Back
Top