Slashdot Mirror
A Brief History Of NVIDIA And SEGA
Alan writes: "FiringSquad just posted an article on the history of NVIDIA. What makes this interesting is that they include a little bit about the NV2 chip which was developed originally for the Dreamcast. It was using quadratic texture maps (a derivative of NURBS) rather than polygons! The article is over here."
The hardware folks like open, certified specs [PCI, AGP, USB, etc.] that they can conform to. Any company with a few talented chip engineers could make a card that outperforms the GeForce 2's, make it plug into any AGP board, and compete.
While to some extent that's true, in practice you'd get your arse sued off until you could prove in court that you really *weren't* violating any of nVidia's implementation patents (or anyone else's). There were a few sabre-rattling sessions last year in this vein.
There are only a few straightforward approaches to building any given part of a graphics pipeline. Just about all of these are patented (usually preemptively) by the big graphics card companies. It's not as bad as the software patent arena, but it's still not nice.
If you have lots of money, you can hold out long enough that the big companies will offer to cross-licence technology with you. Otherwise, you'd better pray that they don't consider you a threat.
Perhaps it's not quite this bad, but you'd have to do quite a bit of patent research to avoid stepping on anyone's toes.
Lastly, a nit-pick re. AGP, PCI, and USB. For some of these, you have to pay licensing fees to get the specification manual. For all, if I understand correctly, you have to pay licensing fees to build any hardware that talks to them. The standards bodies are money-driven too.
In practice, the cost will be low compared to the cost of the rest of your card, but it's still there.
What makes this interesting is that they include a little bit about the NV2 chip which was developed originally for the Dreamcast. It was using quadratic texture maps (a derivative of NURBS) rather than polygons!
I'm sorry but this makes no sense. NV2 chip used qudratic surfaces and not quadratic texture maps. This is like comparing apples and oranges.
Right after a lawsuit by SGI against nVidia over a PCI DMA technique was settled, they allied and nVidia now has a lot of SGI IP at their disposal. The OpenGL stuff they got from SGI is probably why they can't open up the source to their XFree86 drivers.
The article hints that the NV1's quadratic surfaces might have actually been a good thing, and it was held back by Microsoft's push to conformity with triangles.
:-)
Er, no.
For several years now, Nvidia has been kicking ass like no other graphics company, but lets not romanticize the early days. The NV1 sucked bad, and it would have been damaging to the development of 3D accelerators if it had gotten more widespread success. Microsoft did a good thing by standing firm against Nvidia's pressure to add quadratic surfaces to the initial version of D3D.
There is an intuitive notion that curved surfaces are "better" than triangles, because it takes lots of triangles to aproximate a curved surface.
In their most general form, they can be degenerated to perform the same functions as triangles, just at a huge waste in specification traffic.
Unfortunately, there have been a long string of products that miss the "most general form" part, and implement some form of patch surface that requires textures to be aligned with the patch isoparms. This seems to stem from a background in 2D graphics, where the natural progression from sliding sprites around goes to scaling them, then rotating them, then projecting them, then curving them.
3DO did it. Saturn did it. NV1 did it. Some people are probably working on displacement mapping schemes right now that are making the same mistake.
Without the ability to separate the texturing from the geometry, you can't clip any geometry in a general way (not even mentioning the fact that clipping a curve along anything but an isoparm will raise it's order), and you either live with texel density varying wildly and degenerating to points, or you have texture seams between every change in density. No ability to rotate a texture on a surface, project a texture across multiple surfaces, etc. You can't replace the generality if a triangle with primitives like that.
Even aside from the theoretical issues, NV1 didn't have any form of hidden surface removal, and the curve subdivision didn't stitch, frustum clip or do perspective. It was a gimmick, not a tool.
All water under the bridge now, of course. NV20 rocks.
John Carmack
Video card makers actually ASK HIM what he wants to see in the next generation video cards, then scurry off and figure out how to do that. I don't doubt that they send him a few engineering samples once they have working prototypes.
No company is so suicidal that they are going to create a video card that won't run Doom2001 (or whatever its going to be called) Perfectly.
Nvidia never used unfair marketing tactics. They have used some questionable ones when it came to a few websites, but that has been fixed.
Nvidia achieved market dominance by providing good products at good prices, and coming out with new ones so fast to overwhelm their competitors. This is not an unfair business tactic, they are just completely ruthless competitors.
When Nvidia came out with the Riva 128, it had one advantage over the Voodoo. It could do 2D. And everyone had a 2D card, so it really didn't matter. It was slower with the early drivers, it had crap image quality, and it ran Quake 2 at 10fps or so on my P133.
The Voodoo2 came out and Nvidia had nothing but a slightly improved Riva 128, but at this point, people still didn't really care, but the 3d only thing was starting to have an effect. People would run systems with Riva 128s and Voodoo2s for good 2d and great 3d, and the option of real openGL.
With the TNT, Nvidia had a performance competitive product that had much better image quality too, and 3dfx's Banshee was a rehased, weakened Voodoo2. When the Voodoo3 came out, Nvidia promptly took the performance lead with the TNT2/TNT2 Ultra, albeit by a small margin. From then on 3dfx was way too slow to stand a chance. ATI started to come back into the picture here, and they have been tagging along since.
Nvidia releases new products too quickly for others to remain competitive, and they work quickly towards incorporating features OEMs want, lowering their prices, and look at the extremely dominant set of video chips Nvidia has today. Everything they have out is best of class, or damn near it.
MS won the market using completely ruthless tactics, and now they are their own greatest threat. Nvidia is still pumping out products like there is no tomorrow, and they are aware that if they get lax like 3dfx did, they will fall very quickly. ATI, while lagging behind, could jump right back into the front with one botched product release on Nvidia's behalf.
Also, if ATI gets their act together quickly and writes some decent drivers, and gets a comprable card out at a significantly lower price, then ATI will be able to get some of that lost marketshare back. It will take a better deal to beat Nvidia, however, because the established brand can always charge more for the same thing. And ATI has a huge disadvantage there too because they are competing with a ton of smaller boardmakers that can charge whatever they want.
Only those who dream can grasp reality.
What I am supprised to see is the lack of mention in the article to NVidia and SGI. The new SGI Intel boxes are running a variation of the nv15 chipset (and do a good job of it). Also, NVidia, if I remember right, has a ton of ex-SGI employees.
-I just work here... how am I supposed to know?
NVidia is not, and really is not likely to become, a monopolist in any meaningful sense. The hardware side of computers is not like the software. Most specs are open, the closed ones tend not to do well [only one I can recall in recent times is the ZIP, and I think that is due to a decent product, reasonable price, and good customer service when they screw up], just look at apple, or LS120 drives. The hardware folks like open, certified specs [PCI, AGP, USB, etc.] that they can conform to. Any company with a few talented chip engineers could make a card that outperforms the GeForce 2's, make it plug into any AGP board, and compete. The reason this doesn't work in software is because the specs aren't open, they are controlled by the monopolist. So, untill NVidia releases their own slot for graphics boards, patented by them and used soley by their cards, don't get your panties in a bunch over their 'monopoly'. As 3dfx showed us, such monopolies in hardware rarely last long.
-={(Astynax)}=-
-={(Astynax)}=-
"Darkness beyond Twilight"
NVidia does not have a monopoly in any market right now. What they do offer is the best (IMHO) consumer-level graphics card on the market. As well as the second-best. And the third-best. (I'm talking GeForce2 Ultra, GeForce2 Pro, and GeForce 2 here.) That is very different from having a monopoly.
The ATI Radeon is also a reasonably good card, especially if you get the all-in-wonder version. EvilKyro, or whatever it is called, looks interesting, but I don't know all the details about it yet. The point is, NVidia does not have a monopoly.
If you think about it, NVidia is now at the point that 3dfx was at a few years back. A year ahead of the competition. However, unlike 3dfx, NVidia is not encouraging people to use a proprietary API that only works with their own hardware. The best way to access an NVidia card is via OpenGL (even on Windows), which is the most open, cross-platform 3D API starndard out there. NVidia has implemented proprietary extensions to OpenGL, but the extensions are purposely laid out such that it is easy to write software which only uses the extensions if they are available, and uses GL standard calls otherwise. (I know this from experience...)
Furthermore, unlike some well-known monopolistic companies, NVidia is still innovating at an alarming rate. As long as they keep doing so, and their prices stay where they are (GeForce 2 for $130, anyone?), I'm happy.
------
The difference here between Nvidia and microsoft is that Nvidia isn't using monopolistic tactics, and there are viable competitors on the same platform. Nvidia just has better products, and noone else has a better card. Would you say that the Radeon isn't a good card with a nice bit of market share? If you go to any Gateway store (I was there earlier today) and want one of the flat panel monitors, which card do you end up with? Not the Nvidia product, but the 32mb Radeon.
Sorry, but it's really not fair to put them in the same category as Microsoft just because they have the best cards out there that people want.
"Well kids, you tried your best, and you failed. The lesson is, never try." -Homer Simpson
Uhm, sorry but are you dumb? 3dfx had a complete strangehold on the 3D market by locking down performance, price, and reputation. NVidia single-handedly dethroned them using superior products and a better R&D pipeline. You talk like NVidia has always been the big guy on the block when EVERYONE knows they fought a tremendous battle against an undisputed market leader. And won.
---
I had one, on a Diamond Edge2000. It shipped in a huge pile of hype about being:
(a) The first Direct3D accelerator and
(b) Compatible with the Sega Saturn.
Both of which turned out to be bullshit. While it did ship some actually quite good DirectX drivers, the direct 3d aspect of it was being "worked on". They did eventually ship some, with a huge disclaimer along the lines of "we know these things blow chunks, it's all due to this quadratic surface thing". This was all in the middle of John Carmack's big Direct3d rant on usenet ("I am looking forward to doing an apples to apples comparison of Direct3D and OpenGL" turned into "I am not going to finish, and there will never be, a Direct3d port of quake").
It also had onboard sound that was also very damn good. The MPU401 in particular was of near sound canvas quality. Unfortunately it was not sound blaster compatible and since the direct 3d port of quake was never going to happen, games basically stayed in DOS and sound effects had to be sent to a separate card.
Saturn compatibility turned into "Sega will be porting games", which of course they never bothered to do because only eight people bought the cards.
And Diamond were just shit about the whole thing. Haven't bought a thing from them since, don't know anyone else who has either.
So in essence: Great silicon, some serious forward steps were taken; Shitty marketing, I guess they learned.
Dave
BTW, while we're on the subject of bullshit graphics accelerators, did anyone ever get texture mapping going on a Millennium2?
I write a blog now, you should be afraid.
Tiling and the quadratic stuff are completely different. The quadratic stuff is to have a curved primitive. While that's a good thing, most surfaces you'll want to draw won't be quadratic, so you're just slightly better off than if you had triangles. There were also problems in how to do the texturing (well, actually, texturing wasn't too important back then so it didn't go into NV2.) Note that the new DirectX 8 has cubic splines in it, which ought to be better than quadratic primitives.
The modern approach to drawing these curved primitives is to tesselate them into many triangles. If you had to draw a cubic spline using some quadratic primitive, you would have to "tesselate" your cubic primitive into smaller quadratic ones.
Tiling is where you divide your screen into tiles (say, 16x16 pixels per tile). The hardware buffers all your glTriangles (or whatever), doesn't draw a thing until you glSwapBuffers (which would normally display the picture). At that point, it will sort all the triangles into the tiles (some triangles stradle several tiles). Then, each tile is drawn one after the other. The main advantage is that this per-tile work can all be done on-chip instead of doing it in frame buffer memory. Frame buffer bandwidth is one of the most limiting aspects of modern graphics chip design.
The Kyro also uses a scan-line algorithm (if I understand correctly). In particular, it actually sort all the surfaces from front to back (or maybe back to front) to get transparencies correctly (or at least, its Sega incarnation in the DreamCast does this). Unfortunately, this great feature will run against 3d API's (both DirectX and OpenGL) which don't sit easily with that kind of business. Nevertheless, I do believe that such an extension will go into the API's in the near future; the developers are asking for it and the chip designers are trying to get it working.
I design 3d chips for an important graphics company but because of my company's culture, I'd better stay anonymous and not even say whom I work for.