Why hasn't a third vendor come in?

[Anarchist420]

I've been wondering why it's still just ATi and nvidia. Nvidia does more things that I like, but they're not quite perfect (good enough though). They have good filtering, more natural colors, better GL drivers, better depth precision (ATi uses some z-range hack and only offers a 16 bit z-buffer in Shadowman), and extra features including PhysX, plus AA works with more games on nvidia hardware.

Unfortunately, they don't emulate dithering nor do they emulate the w-buffer properly. Neither does ATi. The former isn't too much of a problem because of nglide, but there are still some older dx games that haven't been treated.

 

[Lonyo]

Because it's hard to make a graphics card?
It's not like no one has tried, it's just that they haven't managed to succeed.

Intel have not made it despite kind of twice trying to break the market.
AMD didn't even bother trying (they just bought ATI).
Matrox were a player once but fell out of the market.
Imagination Technologies/PowerVR used to play, but they have gone to integrated mobile/low power stuff.
SiS failed too.
S3.
Trident.
XGI.

 

[Madcatatlas]

Cleverly thought-out ATI bashing.


since you mention older dx games, you must surely have heard of ALL the other gpu producers. Let me refresh it for you, some of them are still making cards. Matrox, 3dfx, S3, Via

 

[KIAman]

Because the market is small compared to onboard OEM GPUs.

 

[exar333]

High start-up costs combined with a competitive market. If the margins where there, Intel would be more focused on it, but they have tepidly worked with GPUs just enough to keep their IGP stuff afloat.

 

[God Mode]

Patents make startup unprofitable.

 

[CurseTheSky]

95 percent (made-up number) of the computer-buying population has no use for a discrete graphics card. The few of us that do either want to play more complex games on the computer than Farmville and Bejeweled, or crunch numbers in a work-oriented environment.

 

[tcsenter]

Why hasn't a third vendor come in?

Loan me $100 million for 10 years at a nominal interest rate, I think we could make a really good go at it. No guarantees, though. You might end-up getting back 5 cents on the dollar.

 

[Phynaz]

Loan me $100 million for 10 years at a nominal interest rate, I think we could make a really good go at it. No guarantees, though. You might end-up getting back 5 cents on the dollar.

Add a zero onto that.

A modern GPU from nothing to a selling product? $1B easy.

 

[sxr7171]

So we had Intel that is probably worth more than AMD and nVidia combined try to make a discrete GPU and they couldn't do it. So who do you propose should come in and take care of business here?

 

[pmv]

There presumably could be an interesting book written about why these two particular companies ended up top-dogs. What did Matrox do wrong, for example? They were in the business before either nvidia or ATI, I believe. The 3dfx story has already been told at web-site length, but I'd be interested to read their rise-and-fall tale as part of such a book.

Presumably though such a book would have trouble telling the real inside-story, dirty-tricks and internal struggles and all, as long as those two corporations are alive and kicking (mostly kicking each other).

 

[erikistired]

i miss 3dfx and matrox.

 

[KIAman]

There presumably could be an interesting book written about why these two particular companies ended up top-dogs. What did Matrox do wrong, for example? They were in the business before either nvidia or ATI, I believe. The 3dfx story has already been told at web-site length, but I'd be interested to read their rise-and-fall tale as part of such a book.

Presumably though such a book would have trouble telling the real inside-story, dirty-tricks and internal struggles and all, as long as those two corporations are alive and kicking (mostly kicking each other).

Easy. 3dfx puchased by nvidia and Matrox moved towards business productivity rather than 3d performance.

 

[SirPauly]

This may be deemed semantic by some but nVidia really purchased the core assets of 3dfx and not the company. Back in the past, there were many competitors, and many tried, but only the ones with vision and talent survived. It's hard what they do.

 

[balane]

I'm stunned you think nVidia has better colors than ATI. I was an nVidia user for over a decade but recently switched to ATI because I thought the shading, lighting and especially the color palette were far superior on the ATI cards. To me it's night and day.

 

[Phynaz]

So we had Intel that is probably worth more than AMD and nVidia combined try to make a discrete GPU and they couldn't do it.

Saying they couldn't do it might be a stretch.

How about their particular x86 everywhere approach didn't work in the GPU space?

With the resources Intel has, if they really wanted to I'm sure they could develop a GPU. They have $18B cash on hand and you can buy a LOT of scientists and engineers with that.

Maybe Intel doesn't see a large enough market for discreet GPUs yet. Maybe they will wait until Nvidia gets squeezed out of the market and pick them up at a fire sale price. Maybe they will never stop thinking of themselves as an x86 CPU company.

 

[digitaldurandal]

A company made an attempt in around 2004, claiming they were going to release scalable multicore cards that would own the competition. Unfortunately they released them without any decent drivers and at a very high price.

The cards wouldn't run half the games that were on the test suites back then. After that the company just gave it up.

 

[pmv]

Easy. 3dfx puchased by nvidia and Matrox moved towards business productivity rather than 3d performance.

Well they were purchased by nvidia _because_ they'd been soundly beaten in the market. That was rather the post-script than the cause. I mean the story has been told before, but it could still make interesting reading with insider accounts of how it all went wrong.

Similarly I bet there were some stressful times at nividia while they tried to get Fermi out the door, which might be interesting to read about from an insiders warts-and-all perspective.

Similarly, I'm not clear _why_ Matrox decided not to compete in 3d performance or why all the rest of them (S3 etc) failed to become players there while ATI and nvidia did.

I just think corporate-wars and technology make an interesting story, that's all.

 

[Idontcare]

bitboys ftw

 

[nemesismk2]

Because the market is small compared to onboard OEM GPUs.

Couldnt of said it better myself

 

[biostud]

because noone will buy the 3rd best product.

 

[sandorski]

There presumably could be an interesting book written about why these two particular companies ended up top-dogs. What did Matrox do wrong, for example? They were in the business before either nvidia or ATI, I believe. The 3dfx story has already been told at web-site length, but I'd be interested to read their rise-and-fall tale as part of such a book.

Presumably though such a book would have trouble telling the real inside-story, dirty-tricks and internal struggles and all, as long as those two corporations are alive and kicking (mostly kicking each other).

I'm pretty sure ATI existed before Matrox, but could be wrong. ATI has existed since the Mid-80's.

Ahh, Matrox first video card 1978. ATI founded 1985.

 

[kalniel]

I don't think the title of the thread is entirely fair on nVidia - they are only just in third place behind Intel and AMD.

 

[Sylvanas]

i miss 3dfx and matrox.

+1

 

[RaistlinZ]

I would splooge my pants if a new Voodoo series card was released that blew away nVidia and ATI offerings. Seriously, there would be much splooge.