In the latest issue of CPU Magazine I wrote an article on Intel's Kung Fu. There is a follow up to this article in the next issue as well. Check it out below:
Nvidia dropped a bombshell recently when it all but erased the line it drew in the sand a little over a year ago. In early 2007, it was pretty clear that Nvidia was no longer going to support Intel in its hybrid graphics for notebooks strategy, nor was it going to support SLI on anything other than its own nForce chip-sets. I weighed in on these developments on my blog (Google “Nvidia gives Intel the finger” and “Nvidia’s quest to become a platform company”). At the time, as I mentioned then, Nvidia saw a huge opportunity to compete against Intel’s lucrative chipset business. So, rather than keeping its enemy close as Sun Tzu recommends, Nvidia tried to fight the giant head to head.
What’s transpiring now is a direct result of that decision. What makes this especially interesting is that Jen-Hsun Huang is easily one of the smartest strategic thinkers in the industry, yet this time he got caught in an almost predictable situation. Perhaps sometimes our successes cloud our judgment—but then again, without risk there can be no reward. Either way, these weren’t “flip of the coin” decisions.
Intel’s kung fu seems to be unstoppable. It seems that any time someone challenges the chipmaker, Intel slowly tears the challenger apart bit by bit until there’s little left but hope. Rumor has it that Intel has been secretly hiring top engineers from the graphics industry, and, combined with other bits of evidence, this leads me to suspect that Intel’s Larrabee initiative is more than just another ho-hum Intel Extreme Graphics core; it’s a serious play to make third-party discrete graphics on a notebook redundant.
AMD’s Puma doesn’t help the situation, either. If AMD can scale it, the company will have created a very competitive platform for the entry-level notebook arena. In the meantime, if it’s everything I think it might be, Intel’s Larrabee will make Centrino king of the world. So what’s next for Nvidia?
I imagine that Nvidia will do whatever it can to make high-performance graphics chips more relevant in the general computing space. If Apple or Microsoft, for example, continue to use the GPU to increase computing performance in their operating systems, this would likely bode well for Nvidia.
I love the concept of visual computing, and the performance that Nvidia has demonstrated by using the graphics processor to accelerate applications that are typically CPU-dependent is astounding. That said, if Intel continues on the path that I think it’s on, the GPU space will have a third major graphics competitor, and of course more competition is great for consumers.
When I was 15, I built an RC helicopter from the ground up. Two months and $2,000 later, I went outside for my maiden flight. I sent my model into the air and wrote it off in less than 45 seconds. I was devastated and swore I would never do it again.
Two weeks ago, I came home with a ready-to-fly RC helicopter under my arm. I spent some time with my kids setting it up, and we took it outside: I crashed it in 30 seconds. This time, however, I rebuilt it, and took it up again. I crashed again, went back to the store and bought a bunch of parts, rebuilt it again, and left it in my garage. I then picked up an RC Flight Simulator for my Blackbird; I’ve been practicing ever since and have become an accomplished flyer.
The moral of the story is you can give up, but don’t ever give up in front of your kids. Well, I guess that’s one of two morals. The other is that you probably shouldn’t give up on Intel as a discrete graphics company based on its previous attempts to get into the market. I believe Larrabee could really shake things up in the graphics market this time around; it seems likely that they will at least offer a compelling alternative to current offerings.
So what was I talking about again? Oh yeah, Intel’s kung fu is tight!
Read More...
Summary only...
