Building Tahiti & The Southern Islands

Now that we’ve had a chance to go over the basis of the Graphics Core Next architecture, let’s talk about the finished products.

Today AMD will be launching Tahiti, the first GPU of the Southern Islands family. Southern Islands will initially be composed of 3 GPUs: Tahiti, Pitcairn, and Cape Verde. Tahiti is the largest and most powerful member of the Southern Islands family, while Pitcairn and Cape Verde get progressively smaller.  AMD has not yet announced the branding or launch dates for Pitcarn and Cape Verde, but it typically takes AMD around 6 months to launch a complete family. As such it’s reasonable to expect that all 3 GPUs will have launched by the end of June although there’s a good likelihood of it happening sooner than that.

All 3 GPUs are based on the GCN architecture, and as family members will have similar features while varying the number of functional units accordingly. Along with the architecture change Southern Islands brings with it a slew of additional features that we’ll get to in the following pages, including Partially Resident Texture (PRT) support, PCIe 3.0, FastHDMI, Direct3D 11.1, and AMD’s fixed-function H.264 encoder, the Video Codec Engine.

But today is all about Tahiti, so let’s get down to business.

As we quickly covered in our introduction, Tahiti is a 4.31B transistor GPU based on the GCN architecture and built on TSMC’s new 28nm High-K process.  Due to TSMC canceling their 32nm process last year AMD has had to wait over 2 years for the next full node rather than taking advantage of the half-node process as they typically do, and as a result the jump from Cayman at 40nm to Tahiti at 28nm is much bigger than with past product launches. Whereas Cayman had 2.64B transistors and a die size of 389mm2, Tahiti has a whopping 63% more transistors than Cayman and yet it’s still smaller, coming in at a slightly more petite 365mm2.


GPU Die Size Comparison

At this point AMD hasn’t provided us with the typical board power values for 7970, but we do know that PowerTune is limited to 250W. In terms of design 7970 is clearly intended to work in similar environments as the 6970, in which case power consumption should be similar to the 6970.

Interestingly enough however we’re hearing that 7970 cards are proving to be very overclockable, which is a good sign for the state of TSMC’s 28nm process, and at the same time a bit distressing. Moore’s Law has continued to hold with respect to transistor density, but the power consumption benefits of using smaller nodes has continued to wane. Having a lot of overclocking headroom means that the 7970 has the potential to be much faster, but it also means that the 7970 (and 28nm GPUs in general) are going to be bottlenecked by power. In which case seeing as how we’re already approaching 300W with single-GPU video cards, the performance gains realized from future fabrication processes would be limited to the ever diminishing returns on power consumption improvements.

Diving deeper into Tahiti, as per the GCN architecture Tahiti’s 2048 SPs are organized into 32 Compute Units. Each of these CUs contains 4 texture units and 4 SIMD units, along with a scalar unit and the appropriate cache and registers. At the 7970’s core clock of 925MHz this puts Tahiti’s theoretical FP32 compute performance at 3.79TFLOPs, while its FP64 performance is ¼ that at 947GFLOPs. As GCN’s FP64 performance can be configured for 1/16, ¼, or ½ its FP32 performance it’s not clear at this time whether the 7970’s ¼ rate was a hardware design decision for Tahiti or a software cap that’s specific to the 7970. However as it’s obvious that Tahiti is destined to end up in a FireStream card we will no doubt find out soon enough.

Meanwhile the frontend/command processor for Tahiti is composed of 2 Asynchronous Command Engines (ACEs) and 2 geometry engines. Just as with Cayman each geometry engine can dispatch 1 triangle per clock, giving Tahiti the same theoretical 2 triangle/clock rate as Cayman. As we’ll see however, in practice Tahiti will be much faster than Cayman here due to efficiency improvements.

Looking beyond the frontend and shader cores, we’ve seen a very interesting reorganization of the rest of the GPU as opposed to Cayman. Keeping in mind that AMD’s diagrams are logical diagrams rather than physical diagrams, the fact that the ROPs on Tahiti are not located near the L2 cache and memory controllers in the diagram is not an error. The ROPs have in fact been partially decoupled from the L2 cache and memory controllers, which is also why there are 8 ROP partitions but only 6 memory controllers. Traditionally the ROPs, L2 cache, and memory controllers have all been tightly integrated as ROP operations are extremely bandwidth intensive, making this a very unusual design for AMD to use.

As it turns out, there’s a very good reason that AMD went this route. ROP operations are extremely bandwidth intensive, so much so that even when pairing up ROPs with memory controllers, the ROPs are often still starved of memory bandwidth. With Cayman AMD was not able to reach their peak theoretical ROP throughput even in synthetic tests, never mind in real-world usage. With Tahiti AMD would need to improve their ROP throughput one way or another to keep pace with future games, but because of the low efficiency of their existing ROPs they didn’t need to add any more ROP hardware, they merely needed to improve the efficiency of what they already had.

The solution to that was rather counter-intuitive: decouple the ROPs from the memory controllers. By servicing the ROPs through a crossbar AMD can hold the number of ROPs constant at 32 while increasing the width of the memory bus by 50%. The end result is that the same number of ROPs perform better by having access to the additional bandwidth they need.

The big question right now, and one we don’t have an answer to, is what were the tradeoffs for decoupling the ROPs? Clearly the crossbar design has improved ROP performance through the amount of memory bandwidth they can access, but did it impact anything else? The most obvious tradeoff here would be for potentially higher latency, but there may be other aspects that we haven’t realized yet.

On that note, let’s discuss the memory controllers quickly. Tahiti’s memory controllers aren’t significantly different from Cayman’s but there are more of them, 50% more in fact, forming a 384bit memory bus. AMD has long shied away from non-power of 2 memory busses, and indeed the last time they even had a memory bus bigger than 256bits was with the ill-fated 2900XT, but at this point in time AMD has already nearly reached the practical limits of GDDR5. AMD’s ROPs needed more memory bandwidth, but even more than that AMD needed more memory bandwidth to ensure Tahiti had competitive compute performance, and as such they had little choice but to widen their memory bus to 384bits wide by adding another 2 memory controllers.

It’s worth noting though that the addition of 2 more memory controllers also improves AMD’s cache situation. With 128KB of L2 cache being tied to each memory controller, the additional controllers gave AMD 768KB of L2 cache, rather than the 512KB that a 256bit memory bus would be paired with.

A Quick Refresher, Cont Tahiti: The First Direct3D 11.1 GPU
Comments Locked

292 Comments

View All Comments

  • haukionkannel - Thursday, December 22, 2011 - link

    Well, 7970 and other GCN based new cards are not so much driver depended as those older radeons. So the improvements are not going to be so great, but surely there will be some! So the gap between 580 or 6970 vs 7970 is going to be wider, but do not expect as big steps as 6970 got via new sets of drivers.
  • Ryan Smith - Thursday, December 22, 2011 - link

    This is actually an excellent point. Drivers will still play a big part in performance, but with GCN the shader compiler in particular is now no longer the end all and be all of shader performance as the CUs can do their own scheduling.
  • CeriseCogburn - Thursday, March 8, 2012 - link

    I hate to say it but once you implement a 10% IQ cheat, it's though to do it again and get away with it again in stock drivers.
    I see the 797x has finally got something to control the excessive shimmering... that's about 5 years of fail finally contained...that I've more or less been told to ignore.... until the 100+ gig zip download here... to prove amd has at least finally dealt with one IQ epic fail... (of course all the reviewers claim there are no differences all the time - after pointing out the 10% cheat, then forgetting about it, having the shimmer, then "not noticing it in game" - etc).
    I'm just GLAD amd finally did something about that particular one of their problems.
    Halleluiah !
    Now some PhysX (fine bullet or open cl but for pete sakes nvidia is also ahead on both of those!) and AA working even when cranking it to 4X plus would be great... hopefully their new arch CAN DO.
    If I get a couple 7970's am I going to regret it is my question - how much still doesn't work and or is inferior to nvidia... I guess I'll learn to ignore it all.
  • IceDread - Thursday, December 22, 2011 - link

    It's a good card, but for me it's not worth it to upgrade from a 5970 to a 7970. Looks like that would be about the same performance.
  • Scali - Thursday, December 22, 2011 - link

    This is exactly the reason why I made Endless City available for Radeons:
    http://scalibq.wordpress.com/2010/11/25/running-nv...

    Could you run it and give some framerate numbers with FRAPS or such?
  • Boissez - Thursday, December 22, 2011 - link

    What many seem to be missing is that it is actually CHEAPER than the current street prices on the 3GB-equiped GTX 580. IOW it offers superior performance, features, thermals, etc. at a lower price than current gen at a lower price.

    What AMD should do is get a 1.5 GB model out @450$ ASAP.
  • SlyNine - Thursday, December 22, 2011 - link

    Looks like I'll be sticking with my 5870. I upgraded from 2 8800GT's ( that in SLI never functioned quite right because they were hitting over 100C ever with after market HSF) and enjoyed over 2x the performance.

    When I upgraded from a 1900XT to the 8800GT's same thing, 800XT-1900XT, 9700pro - 800XT, 4200(nvidia)-9700pro. The list goes on to my first Geforce 256 card.

    Whats the point, My 5870 is 2! generations behind the 7970 yet this would be the worst $per increase in performance yet. Bummer I really want something to drive a new 120hz monitor, if I ever get one. But then thats kinda dependent on whether or not a single GPU can push it.
  • Finally - Thursday, December 22, 2011 - link

    Since when do top-of-the-line cards give you the best FPS/$?
    For the last few months the HD6870+HD6850 were leading all those comparisons by quite some margin. The DH7970 will not change that.
  • SlyNine - Thursday, December 22, 2011 - link

    If you read my post, you will notice that I'm compairing it to the improvments I have paid for in the past.

    40-60% Better than a 2 YO 5870 Is much worse than I have seen so far. Considering that its not just one generation but 2 generations beyond and for 500+$ to boot. This is the worst upgrade for the cost I have seen.....
  • SlyNine - Thursday, December 22, 2011 - link

    The 6870 would not lead the cost per upgrade in performance at all, It would be in the negitives for me.

Log in

Don't have an account? Sign up now