The RV770 Story: Documenting ATI's Road to Success

Final Words

It took Derek and I six thousand, four hundred and sixty one words to review the Radeon HD 4850 and 4870. At this point I’m at 7,788 and all I’ve done is document the gravity of the decision that lead to the RV770.

There’s a lot of work that goes into all of these products we review, both good and bad. These engineers put their life’s work into every last design they complete, both the good ones and the bad ones. To live in the minds of ATI’s engineers as the first R600 reviews were hitting the web is something I would pay anything to avoid.

The life of a chip architect can be quite difficult, to work on something for three years only to have a few poor decisions make it the web’s punching bag is beyond rough. If I screw up a review I can always try to do better next week, if a chip designer contributes to a billion transistor GPU that’s a failure in the market, he/she won’t have another chance to succeed for several months if not a couple of years. I wonder if these chip companies offer counseling as a part of their benefits packages.

There are thousands of stories behind every chip launch, good or bad, most of them never get told. Part of it is that we’re spending so much time praising or berating the product that we rarely have time to offer the backstory. There’s also the issue with most companies being unwilling to disclose information, for any chip company to give me the level of detail that Carrell offered was a big deal, for that chip company to be ATI/AMD is impressive.

We all have these folks to thank, the engineers I met with and the many more that I didn’t. NVIDIA may not have been happiest with the efforts of the RV770 team, but we all benefitted. If you ended up buying a Radeon HD 4800 or derivative, you already know why you’re thankful. If you ended up buying something green, you most likely paid a much lower price than you would have.

It’s often said that competition is good for the market, but rarely do we have such a clear example of it as what happened after the RV770 launch. Cards that used to cost $300 now cost $200, a brand new GPU that was priced at $400 all of the sudden became reasonable at $300. The consumer won; the RV770 team targeted the Performance segment and did a bang up job of addressing its needs.

And it all started because a few guys were willing to shake things up back in 2005.