The Last Days at 3dfx

sand writes "FiringSquad has a detailed account of what happened in the final days at 3dfx. Every 3dfx product that was released or upcoming is discussed by a former 3dfx employee with inside knowledge on what caused the product delays (including an employee who forgot to fly to Asia to pickup the first Voodoo5 chips). He also discusses money mismanagement and the STB merger. It's a very enlightening article for anyone who's interested in 3D graphics and what goes on inside these companies."

Comments

Competition

[e8johan]

This is what competition is all about. When a company cannot deliver the best product to the best price they don't get any income. If you don't have and income and spend alot without being able to overtake your competators, you will enventually run out of money. It is not fun, but reality in a market economy.

Eventually we will see this when it comes to ATI and nVidia, or they will find a niche market to survive in. The big profit will go to the one making the best product at the best price.

Note - I do not critisize market economy, without it we would probably not have hardware accelerated 3D for home computers at all!

Re:Competition

[dpilot]

You've missed the point. There is not supposed to be a Final Winner, or else we all lose. As soon as there is a Final Winner, competition dies. At any one generation, the Winner gets better profits and the right to compete on the next generation. But you want there to be a runner-up who will viably go on to the next generation, as well.

We have let Microsoft color our thinking too much, fill us with envy, and convince us that this is The Business Model.

They didn't innovate enough

[91degrees]

They should have added useful features and clever thinking that circumvented the problems that plagued the other companies. AGP Texture bandwidth could have been solved by texture compression, but S3 ended up doing that. 32 bit colour was implemented by everyone except 3DFX. They could have saved a lot fo bandwidth if they'd have come up with better Z buffer algorithms, but PowerVR did that. They could have added programmable graphics, but that was left to ATI. They could have put T&L on the card, but that was left to Nvidia.

3DFX failed because they didn't innovate

Re:They didn't innovate enough

[CoolVibe]

3DFX failed because they didn't innovate

No, it failed because of braindead and utterly stupid upper management. Most companies die that way.

It's always the management that screws it up. Remember that. Read Dilbert. Understand it. Make it your corporate religion. Prevent falling on your face. Oh, and don't forget: laugh.

Re:They didn't innovate enough

[Junks Jerzey]

3DFX failed because they didn't innovate

That's silly. 3dfx innovated like crazy:

* First high performance consumer level 3D card for PCs.
* First multitexturing in a PC card.
* First trilinear filtering in a PC card.
* Glide API, back when Direct3D and OpenGL were poorly supported on the PC.
* Higher precision color blending with 16 bits per pixel. Operations occurred internally at 22-bits, I think.
* Able to connect multiple Voodoo 2's together for--what was then--unheard of performance.

Let's not rewrite history to fit your own ideas, okay?

Voodoo cards

[AliasTheRoot]

At the start of the consumer 3d graphics business Voodoo were clearly superior, I still have a Voodoo 1 laying around somewhere, there were problems; the whole passthrough cable thing, the lack of windowed support & 16bit clour were all problematic. As an upgrade Voodoo offered the second revision that could run in SLI mode. It required two PCI slots in addition to your 2d graphics card and was horrendously expensive.

nVidia released the TNT that offered similar performance, in one card (not 3!), did 32 bit colour and was significantly cheaper.

3DFX was never competitive from then on, offering weaker, more expensive products that relied on brand name support.

The widespread adoption of D3D / OpenGL around this time over the proprietary Glide API was the nail in the coffin.

GLIDE

[muzzmac]

I like many others was not concerned with them going. Thier attempt to lock the market in via the proprietray GLIDE API was a blatant move to control the market.

I'm happy to see the tail end of any company that does this.

Thier lawsuit against the guy doing the GLIDE wrapper didn't help improve my opinion of them. :-)

Re:GLIDE

[Sycraft-fu]

Actually Glide orignally was a benign move. Back in the day when the Voodoo first was being realised, there weren't any acceptable APIs out there. DirectX was still too primitive to be really taken seriously and OpenGL was just more than the Voodoo could handle quickly. Hence Glide, which was optimised for how the Voodoo hardware worked.

The first Voodoo really was a pretty amazing hack to make it work at all. When 3dfx first demoed their new card on a simulator, they got laughed at, people said they'd never make it real silicon. It was therough a lot of ingenuity and scaling back features that they managed to build a 3d card at a consumer pricepoint. It was expensive, yes, but not the thousands of dollars pro cards cost.

Their big problem later was that they really failed to move forward. Technology progressed to the point where you didn't have to make all the compramises and cards like the TNT and TNT2 proved it. Also, Glide was a relic that they should have tried to phase out since DirectX did come to mature and cards had no trouble with OpenGL.

Sure!

[zrk]

Most secret technology is often sent through methods that can be intercepted by halfway decent corporate spies.