Sign in with
Sign up | Sign in

Building Cayman By Improving Cypress

Radeon HD 6970 And 6950 Review: Is Cayman A Gator Or A Crock?
By

According to AMD, it had four principal design goals in building Cayman: more efficiency, improved geometry performance, new image quality features, and better power management.

First, it wanted to create a more efficient graphics and compute architecture. The motivation behind this decision is sound enough—AMD was seeing a VLIW rate of roughly 3.4 in games. So, removing the special function transcendental unit and distributing its functionality across the other four units was actually a good performance per area optimization that promised to keep the GPU running within the observed operating rate. There are situations where performance could take a hit (when the VLIW utilization spikes above four), but AMD says that's unlikely.

More important, AMD needed to create a more efficient architecture. Stuck on TSMC’s 40 nm manufacturing node, the company had to figure how to get more performance per millimeter of die space, rather than simply focusing on adding absolute performance. By shifting from its five-way VLIW architecture to a four-way design, AMD claims a 10% improvement to performance per square millimeter of die, as it’s able to add more SIMDs to the same amount of space.

Cypress/Barts employs a VLIW5 architectureCypress/Barts employs a VLIW5 architecture

Cayman employs a more efficient/area VLIW4 configurationCayman employs a more efficient/area VLIW4 configuration

Streamlining the architecture doesn’t make it any less capable. The four stream processors now have identical capabilities, absorbing the special function unit’s role as well. In its VLIW4 configuration, each stream processor can do:

  • Four 32-bit FP FMA, MAD, MUL, or ADD per clock
  • Two 64-bit FP ADD per clock
  • One 64-bit FP FMA or MUL per clock
  • One FP Special Function per clock
  • Four 24-bit Int MAD, MUL, or ADD per clock
  • Four 32-bit Int ADD or bitwise opps per clock
  • One 32-bit Int MAD or MUL per clock
  • One 64-bit ADD per clock


Moving beyond the GPU’s shading core, its render back-ends are able to handle 16-bit integer ops 2x faster, while 32-bit FP ops are 2x-4x faster. According to AMD, this most directly affects anti-aliasing performance.

Augmenting Compute Performance

Although AMD’s compute-oriented aspirations are often taken less seriously than Nvidia’s, this does sound like an area that received some attention with Cayman. For instance, whereas the Radeon HD 5800-series cards perform double-precision math at one-fifth of the single-precision rate, Cayman operates at one-quarter the SP rate. Although the Radeon HD 6970’s peak single-precision rate is a touch lower than Radeon HD 5870 (2.7 TFLOPS versus 2.72 TFLOPS), you end up with 675 GFLOPS of peak double-precision math on the Radeon HD 6970 compared to 5870’s 544 GFLOPS.

Note also that the Barts GPU sacrifices DP altogether, focusing on gaming performance rather than compute capabilities.

Cayman also incorporates dual bidirectional DMA engines, which ideally yield faster reads and writes to and from system memory over the PCI Express bus.

Finally, AMD gives Cayman the ability to handle independent applications across the GPU. This is in contrast to Fermi, which can handle multiple kernels, so long as they’re spawned from the same CPU thread. Interestingly, that functionality isn’t part of DirectX 11, so AMD has to instead enable it through OpenCL sometime in the future.

Aside from those functionality tweaks, Cayman retains Cypress’ cache structure. Each SIMD has its own 8 KB L1 cache for computational work, aside from the 16 KB L1 texture cache, plus a 32 KB local data share. Four 128 KB L2 caches continue keeping those SIMDs fed with information, and there is still a 64 KB global repository shared by all of the SIMDs.

Ask a Category Expert

Create a new thread in the Reviews comments forum about this subject

Example: Notebook, Android, SSD hard drive

Display all 217 comments.
This thread is closed for comments
Top Comments
  • 30 Hide
    terror112 , December 15, 2010 3:13 AM
    WOW not impressed.
  • 15 Hide
    iam2thecrowe , December 15, 2010 4:22 AM
    Dissappointed. well i guess anyone that bought a 5xxx series card that couldnt wait can now be content that they made an ok choice. The only thing i got from this review is that i want 2 x gtx460's or 2 x 6850's, not the new amd cards.
  • 14 Hide
    bluekoala , December 15, 2010 4:56 AM
    I also urge people to put more emphasis on MINIMUM FPS as that is when you require high FPS the most.
Other Comments
  • 30 Hide
    terror112 , December 15, 2010 3:13 AM
    WOW not impressed.
  • 11 Hide
    Annisman , December 15, 2010 3:15 AM
    Thanks for the review Angelini, these new naming schemes are hurting my head, sometimes the only way to tell (at a quick glance) which AMD card matches up to what Nvidia card, is by comparing the prices, which I think is bad for the average consumer.
  • 13 Hide
    rohitbaran , December 15, 2010 3:25 AM
    These cards are to GTX 500 series what 4000 series was to GTX 200. Not the fastest at their time but offer killer performance and feature set for the price. I too expected 6900 to be close to GTX 580, but it didn't turn out that way. Still, it is the card I have waited for to upgrade. Right in my budget.
  • 3 Hide
    tacoslave , December 15, 2010 3:25 AM
    imagine when this hits 32nm?
  • 7 Hide
    notty22 , December 15, 2010 3:27 AM
    AMD's top card is about a draw with the gtx 570.
    Pricing is in line.
    Gives AMD only hold outs buying options, Nvidia already offered
    Merry Christmas
  • 11 Hide
    microterf , December 15, 2010 3:27 AM
    Why drop the 580 when it comes to the multi-gpu scaling??
  • 4 Hide
    IzzyCraft , December 15, 2010 3:28 AM
    Sorry all i read was this
    "This helps catch AMD up to Nvidia. However, Intel has something waiting in the wings that’ll take both graphics companies by surprise. In a couple of weeks, we'll be able to tell you more." and now i'm fixated to weather or not intel's gpu's can actually commit to proper playback.
  • -3 Hide
    andrewcutter , December 15, 2010 3:28 AM
    but from what i read at hardocp, though it is priced alongside the 570, 6970 was benched against the 580 and they were trading blows... So toms has it at par with 570 but hard has it on par with 580.. now im confused because if it can give 580 perfomance or almost 580 performance at 570 price and power then this one is a winner. Sim a 6950 was trading blows with 570 there. So i am very confused
  • -1 Hide
    sgt bombulous , December 15, 2010 3:30 AM
    This is hilarious... How long ago was it that there were ATI fanboys blabbering "The 6970 is gonna be 80% faster than the GTX 580!!!". And then reality hit...
  • 2 Hide
    manitoublack , December 15, 2010 3:35 AM
    I'd have to say wait until the christmas new years dust settles
  • 8 Hide
    andrewcutter , December 15, 2010 3:36 AM
    sry i take back what i said earlier. most reviews agree with what toms says. So my apologies..:) 
  • 6 Hide
    tpi2007 , December 15, 2010 3:36 AM
    Not bad, but not very impressive either. It's hard to be impressed at 40nm by now.

    But it is quite ironic that AMD has had a tesselator in their cards way before anybody supported the feature (let alone Nvidia), and now Nvidia does better tessellation than AMD.. they should really address that problem.. well, now the only way is to redesign the chip... at 28nm.

    28nm it is then, the next big excitment.

    What I would really like, now that the HD6xxx lineup is here (dual GPU still missing, but that is a niche product), is that AMD now focuses on fixing bugs in their drivers.
  • 4 Hide
    namelessonez , December 15, 2010 3:39 AM
    It's always the same story! nVidia pulls out a new product and then the wait begins for AMD to release its products! Ultimately, the difference isn't worth the wait. We know that AMD's winning factor is its price, but nVidia's is the quality....imho!

    As rightly stated, 'reality hits'.
  • 0 Hide
    yyk71200 , December 15, 2010 3:50 AM
    AndrewCutterbut from what i read at hardocp, though it is priced alongside the 570, 6970 was benched against the 580 and they were trading blows... So toms has it at par with 570 but hard has it on par with 580.. now im confused because if it can give 580 perfomance or almost 580 performance at 570 price and power then this one is a winner. Sim a 6950 was trading blows with 570 there. So i am very confused

    Actually, in Hardocp review overall 580 has some edge over 6970 as well. Only in F1 6970 is ahead. 6970 is great value though.
  • 8 Hide
    joytech22 , December 15, 2010 3:52 AM
    IzzyCraftSorry all i read was this"This helps catch AMD up to Nvidia. However, Intel has something waiting in the wings that’ll take both graphics companies by surprise. In a couple of weeks, we'll be able to tell you more." and now i'm fixated to weather or not intel's gpu's can actually commit to proper playback.


    If intel entered the graphics market and provided a half-decent dedicated GPU, that would definitely make ANY GPU company shake in their boots.

    But in all honesty i hope Intel does enter the market for graphics, making AMD and Nvidia push harder and faster for better products.
  • 1 Hide
    Stardude82 , December 15, 2010 3:53 AM
    So the answer to the question posed in the title is that it is neither really a gator nor a crock. It works on so many levels! Well, maybe just one since caimans are more closely related to alligators than crocodiles.
  • 10 Hide
    Tamz_msc , December 15, 2010 3:54 AM
    I wonder if Nvidia will reduce the prices on the GTX 580 and 570.
  • 6 Hide
    Lamiel , December 15, 2010 4:02 AM
    About the only good news I can see in this for AMD is how much they've increased their multi-GPU scaling, making the 6850's in Crossfire a great bargain. That's great, sure, but the new 6900 cards leave me completely underwhelmed. It doesn't make a whole lot of sense to follow up the 6800's in this way. I'm an Nvidia user, but I was still looking forward with curiosity to seeing how much stronger the 6970 would be than the GTX 580. Looks like the hype-machine broke down this time... My guess is that a lot of the AMD/ATI fanatics will be scrambling to salvage some dignity after all their talking up of Cayman and how it would eat Nvidia's lunch.
  • -3 Hide
    fstrthnu , December 15, 2010 4:16 AM
    This reminds me a lot about the recent release of the updated AMD processors - a temporary holdout while the company plans to release new ones ahead. Except these graphics cards don't even have value going for them. Nvidia was behind just a couple months ago, but now they're ready to crush AMD/Radeon. AMD better have something new coming, and FAST. Holding procedures will only do so much when your competitor is already developing their next-gen GPUs.
  • 15 Hide
    iam2thecrowe , December 15, 2010 4:22 AM
    Dissappointed. well i guess anyone that bought a 5xxx series card that couldnt wait can now be content that they made an ok choice. The only thing i got from this review is that i want 2 x gtx460's or 2 x 6850's, not the new amd cards.
Display more comments