The RV770 Story: Documenting ATI's Road to Success
by Anand Lal Shimpi on December 2, 2008 12:00 AM EST- Posted in
- GPUs
Just One Small Problem: We Need a New Memory Technology
The R600 GPU had an incredibly wide 512-bit memory interface, the problem with such a large interface is that it artificially makes your die bigger as you’ve got to route those interface pads to the memory devices on the board. For RV770 to have the die size ATI wanted, it needed to have a 256-bit memory interface, but using (at the time) current memory technology that wouldn’t give the GPU enough memory bandwidth to hit the performance targets ATI wanted.
When the options were either make the chip too big or make the performance too low, ATI looked elsewhere: let’s use a new memory technology. Again, put yourself in ATI’s shoes, the time was 2005 and ATI had just decided to completely throw away the past few years of how-to-win-the-GPU-race and on top of that, even if the strategy were to succeed it would depend on a memory technology that hadn't even been prototyped yet.
The spec wasn’t finalized for GDDR5 at the time, there were no test devices, no interface design, nothing. Just an idea that at some point, there would be memory that could offer twice the bandwidth per pin of GDDR3, which would give ATI the bandwidth of a 512-bit bus, but with a physical 256-bit bus. It’s exactly what ATI needed, so it’s exactly what ATI decided to go with.
Unfortunately whether or not GDDR5 shipped by the summer of 2008 wasn’t all up to ATI, the memory manufacturers themselves had a lot of work to do. ATI committed a lot of resources both monetarily and engineering to working with its memory partners to make sure that not only was the spec ready, but that memory was ready, performing well and available by the summer of 2008. Note that the RV770 was going to be the only GPU that would use GDDR5, meaning that it was ATI and ATI alone driving the accelerated roadmap for this memory technology. It’s akin to you trying to single handedly bring 100Mbps internet to your city; it’ll happen eventually, but if you want it done on your time table you’re going to have to pickup a shovel and start burying a lot of your own cable.
ATI did much of the heavy lifting with the move to GDDR5, and it was risky because even if RV770 worked out perfectly but the memory wasn’t ready in time the GPU would get delayed. RV770 was married to GDDR5 memory, there was no other option, if in three years GDDR5 didn’t ship or had problems, then ATI would not only have no high end GPU, but it would have no performance GPU to sell into the market.
If GDDR5 did work out, then it meant that RV770 could succeed and that it would be another thing that NVIDIA didn’t have at launch. That is, of course, assuming that ATI’s smaller-die strategy would actually work...
Facebook
Twitter
RSS
116 Comments
View All Comments
Spivonious - Wednesday, December 3, 2008 - link
I totally agree! Articles like this one are what separates Anandtech from the multitude of other tech websites.goinginstyle - Wednesday, December 3, 2008 - link
I have to admit this is one of the best articles I have read anywhere on the web in a long time. It is very insightful, interesting, and even compelling at times. Can you do a follow up, only from an NVIDIA perspective.Jorgisven - Wednesday, December 3, 2008 - link
I totally agree. This article is superbly written. One of the best tech articles I've read in a long long time, out of any source, magazine or online. I highly doubt nVidia will be as willing to expose their faults as easily as ATI was to expose their success; but I could be entirely mistaken on that.In either case, well done Anand. And well done ATI! Snagged the HD4850 two days after release during the 25% off Visiontek blunder from Best Buy during release week. I've been happy with it since and can still kick around the 8800GT performance like yesterday's news.
JonnyDough - Wednesday, December 3, 2008 - link
I agree about the insight especially. Gave us a real look at the decision making behind the chips.This got me excited about graphics again, and it leaves me eager to see what will happen in the coming years. This kind of article is what will draw readers back. Thank you Anandtech and the red team for this amazing back stage pass.
magreen - Wednesday, December 3, 2008 - link
Great article! Really compelling story, too.Thanks AMD/ATI for making this possible!
And thanks Anand for continually being the best on the web.
JPForums - Wednesday, December 3, 2008 - link
Like others have said, this is probably the best article I've read in recent memory. It was IMHO well written and interesting. Kudos to ATI as well for divulging the information.I second the notion that similar articles from nVidia and Intel would also be interesting. Any chance of AMD's CPU division doing something similar? I always find the architectural articles interesting, but they gain more significance when you understand the reasoning behind the design.
jordanclock - Wednesday, December 3, 2008 - link
This is easily one of my favorite articles on this website. It really puts a lot of aspects of the GPU design process into perspective, such as the shear amount of time it takes to design one.I also think this article really adds a great deal of humanity to GPU design. The designers of these marvels of technology are often forgotten (if ever known by most) and to hear the story of one of the most successful architectures to date, from the people that fought for this radical departure... It's amazing, to say the least.
I really envy you, Anand. You get to meet the geek world's superheroes.
pattycake0147 - Wednesday, December 3, 2008 - link
I couldn't agree more! This could be the best article I've read here at anandtech period. The performance reviews are great, but once in a while you need something different or refreshing and this is just precisely that.LordanSS - Wednesday, December 3, 2008 - link
Yep, I agree with that. This is simply one of the best articles I've read here.Awesome work, Anand.
Clauzii - Wednesday, December 3, 2008 - link
I totally agree.