The Nvidia Story

While attending Nvidia’s GTC 2013 conference in San Jose I ran into Jen-Hsun Huang the founder of Nvidia.  It’s been many years since we spoke but it brought back a lot of memories of my first encounters with Nvidia when they were first starting out back in the early 1990’s.  I generally don’t feel like I’m adding any value to a conversation when I have positive things to say about anything but I want to put my marker in the ground as saying WOW Nvidia is really killing it these days, it’s amazing what they’ve accomplished since their earliest days and in spite of some tremendous obstacles and a lot of competitors.

http://venturebeat.com/2013/01/06/nvidia-unveils-project-shield-an-awesome-mobile-game-console/

http://venturebeat.com/2013/01/06/nvidia-demos-dead-trigger-2-game-running-on-tegra-4-video/

http://venturebeat.com/2013/01/06/nvidia-launches-its-long-awaited-tegra-4-mobile-processor-for-blazing-fast-tablets/

http://venturebeat.com/2013/01/06/nvidia-unveils-grid-gaming-system-to-jumpstart-cloud-gaming/

I think they are all over the right ideas in gaming lately.  I wonder if the folks making the next generation of consoles, shortly to be announced, realize that they should simply be mobile devices that can plug into a TV as Nvidia and Apple, for that matter, have demonstrated.  I don’t get excited about technology products easily anymore but I’m embarrassed to admit that I’m totally in love with my GeForce 690 dual processor.  I’ve never had so much graphics or computing power at my fingertips.  Sometimes I write code in CUDA just to see what will slow it down.  It’s not easy… but I’m very creative…

Jen-Hsun Huang holding a next generation Nvidia GPU in his fingertips. Apparently it’s too small to be seen by the naked eye, but it features over a trillion transistors.

It’s been many years but I remember when Nvidia was just getting started.  Jen-Hsun came to Microsoft around 1995 when we were in the early stages of designing Direct3D with his then, VP of Marketing, Kevin Dallas.  He had a deal with Sega back then to provide the graphics processor to the next generation Sega Dreamcast.  (The deal didn’t work out for you history buffs who know that the Dreamcast never shipped with an Nvidia chip)  Eric Engstrom and I met with him.  I recall that Jen-Hsun showed us a remarkable product for that era.  A multifunction media card for Windows that supported 3D graphs, 2D graphics, sound and digital joystick on the same card.  It was the first of its kind for the PC.  Jen-Hsun was adamant that the future of 3D graphics would be defined by his Quad based hardware architecture in which 3D objects were rendered as patches of 4 vectors rather than the widely adopted 3 vector based polygon solutions that are the standard today.  Looking back I suspect that might have been a function of their Sega relationship because Sega’s arcade machines were Quad based back then.

I recall that we were very impressed by his product and secretly threatened that he might succeed at pushing Quad based hardware into the PC market. We were our usual pompous Microsoft selves in dealing with Jen-Hsun and told him that we had no plans to support Quad based graphics in Direct3D then subsequently recruited away his VP of Marketing who we were also impressed with.  We proceeded to align ourselves strongly with video chip makers like ATI, Cirrus, Renderware, Matrox and others who were eager to support OUR vision of how 3D on the PC should evolve.  3DFX was the clear leader in consumer hardware in that era and were busy promoting their own proprietary 3D API and driver architecture called Glide which was far more of a competitive threat to Direct3D than OpenGL was at that time because it had real demand and support from the game community and of course 3DFX.

When Direct3D launched we promoted the traditional Windows based chip makers who had worked closely with us and Nvidia’s first generation PC product, with poor DirectX support, was not a commercial success.

I don’t know what was going on in Nvidia’s relationship with Sega at the time but suffice it to say that it wasn’t a done deal.  Another 3D chip company called PowerVR told us THEY had gotten the Sega deal and that Nvidia was being dumped for them.  Microsoft didn’t have a relationship with Sega in the US at that time but soon afterword’s we would, which is a story for another blog.  The short of it is that for a time it seemed that Nvidia was destined to fade into the history books with 3DFX and many other earlier graphics chip makers.  ATI working closely with Microsoft on DirectX hardware and support looked like they might ultimately dominate the space.  Nvidia’s early influence on DirectX was mostly to show us what great integrated consumer media cards for the PC could look like which inspired us to see DirectX as a family of driver API’s that could work across a disparate set of specialized media cards or be a unified driver model for multifunction cards like Nvidia’s in the same architecture.

Unlike their predecessors however, Nivida survived and came back to Microsoft with an entirely different approach.  Jen-Hsun’s new strategy was to be the BEST Windows 3D implementation in the market.  They aggressively adopted DirectX API’s and adapted their hardware to perform extremely well at it.  They hired relationship people to work closely with Microsoft and the DirectX team.  Their new products ran great on our platform and we eagerly promoted them.  The trouble for Nvidia was that something else had changed at Microsoft in the intervening year.  The success of DirectX had caused OTHER groups at Microsoft to have their own ideas about how Microsoft should define and control the consumer game market on the PC.  The other group was the Advanced Technology Group run by Bill Gate’s right hand technologist Nathan Myhrvold.  (Recall in this context that I and the other DirectX folks had been recruited by and worked for Nathans brother Cameron who ran the Developer Relations Group)

Nathan had rounded up some of the most famous luminaries in the 3D world including Jim Kajiya, James Blinn and many others that had taught the 3D courses at Siggraph that I had attended in the 1980’s.  Working for Nathan THEY had cooked up their own 3D hardware architecture called Talisman which they planned to license to Windows chip makers.  The success of Direct3D was an inconvenience for that initiative because Direct3D hardware looked nothing like the “brilliant” Talisman chip design.  The solution to them was obvious; they should own DirectX so that they could use it to foist Talisman off on the hapless chip makers we had talked into supporting DirectX!  DirectX was taken from our control and handed over to the Talisman team, then run by Jay Torborg.  Eric, Craig and I were beside ourselves over this outcome for all our hard work, they moved on to other things at Microsoft while I reluctantly remained to try to minimize the damage.  While the OpenGL faction at Microsoft from the Windows NT team was running their own PR campaign outside Microsoft to promote their control of 3D, DirectX had been handed to a research team making a 3D chip that didn’t resemble any known 3D API or architecture.

http://en.wikipedia.org/wiki/Microsoft_Talisman

I was pressured to use my relationships with the OEMs and developers to adopt Talisman and as this Blog already recounts, my reaction was to hit the red button and go down in a hail of bullets fighting it instead, much more fun.

Looking back the strangest thing about this time was that there where many situations in which I would have sworn that we were failing and doomed when we were actually winning and didn’t know it… and inversely many situations in which I was absolutely certain we were winning when we were actually quite doomed…  When I left Microsoft, I thought it had all been a huge failure.  By succeeding at doing anything with media and gaming at Microsoft I had simply created an enormous internal brawl for control over that success which had spilled chaotically into the industry I had been trying to build.  By succeeding at selling the ideas and building trust in Microsoft with an industry that hadn’t trusted Microsoft before, I had exposed them all to the exact treatment from Microsoft that they had previously feared.  Worst of all, I’d had the technology product that had been built by people who were deeply passionate about it handed to a bunch of research drones who wanted to usurp it to foist their disastrous 3D hardware off on the world and my previous allies had either signed up for it OR were demanding OpenGL support.

…and yet in spite of all of this it still somehow worked out.  To everyone’s GREAT benefit, Nvidia resisted Microsoft’s pressure to adopt Talisman, which was an incredible gamble considering their earlier experience with not aligning with a Microsoft initiative.  In my absence those “idiots” from the Talisman group actually picked up the fallen DirectX banner and ditched their own baby to run with it.  They merged with the OpenGL team and produced several subsequent generations of DirectX that exceeded my greatest hopes for it.  Gates, who I had become extremely disappointed with for his highly misplaced support of Talisman advocated hiring some of the game industry luminaries I had introduced him to including Seamus Blackley.  Seamus and my successor Kevin Bacchus who I recruited from Mindscape started the DirectXBox project.  Microsoft for the first time started recruiting real gamers into their studio and technology teams.  With the arrival of people like Stuart Moulder into the new games group run by Microsoft veteran Ed Fries came games like Age of Empires and Halo.

The first DirectXBox shipped with an Nvidia based architecture.  Since then, for reasons most likely related to cost, Microsoft chose cheaper components for the Xbox 360 making it, regrettably in my opinion, less like a PC and therefore more difficult to create games for that could run well on both platforms which was the original strategy for WHY Microsoft should enter the console business in the first place.  I think that Nvidia’s vision for the future of gaming is the right one and I’m very excited to be alive in an era when I can work with such amazing computing power.  I feel like I lived to see the era when I could walk on the bridge of the Enterprise and play with warp cores…. literally… because that’s what Nvidia calls the smallest unit of parallel threads you can run on a GPU.

Comments

comments