AMD 6970 Crossfire

[Charlton]

During the final presentation of the Radeon HD 6900, AMD announced the color directly the GeForce GTX 580 is out of reach. This will disappoint to some because many thought they saw ahead of AMD with Nvidia GPU performance. AMD also believed but probably also had its sights lower. The cause may be a new architecture which has proved more complex than expected to get the best output but also excellent GeForce GTX 500 which has thwarted the original plans.

With Cayman, AMD took a risk by deciding to review one aspect of the architecture of its computation units that had not changed since the Radeon HD 2900 XT. In a simplified manner, we characterize the computational units of AMD vec5, which means they are capable of executing up to five instructions in parallel. However, with such an architecture, if the code to run does not parallelize as many statements, they will not be fully exploited, in contrast to the scalar architecture from Nvidia that can maintain high efficiency over a maximum of situations. Both approaches equally valid as the other.

Please do not confuse arithmetic unit with Core, a marketing concept used by Nvidia to be compared to CPUs and AMD followed by the opportunity to have 5 cores per unit calculation vec5. Overall you can see things from two angles: one unit vec5 AMD is more efficient than scalar unit from Nvidia or AMD's core is less efficient than Nvidia core. With the GeForce GTX 460 GF104 and its derivatives, Nvidia got closer to a vector operation to increase efficiency and AMD intends to do the same with the Cayman, but in the other direction, from vec5 to vec4 . The calculation units of Cayman are thus less powerful than previous AMD GPUs, but they are statistically more efficient but not more powerful, the distinction is important. Cons by being simple, these units occupy less space and consume less, thereby increasing the number, all other things being equal.

In more detail, the previous Radeon GPUs were based on computing units of type 4 + 1, with a run line can handle complex instructions. This one that AMD has decided to get rid of. In return for these complex instructions to be processed on other lines through a succession of simpler operations. These instructions monopolize and 3 of the 4 lines of execution, making them much more intensive since only a single instruction can be executed in parallel against 4 before. Without this one a bit special and in some cases difficult to feed properly, the compiler will see his task greatly simplified, which in some cases may even make these units more efficient than vec4 previous vec5 but overall AMD now needs more computing units vec4 to maintain the same level of performance.

While Cypress GPU Radeon HD 5800, had 20 blocks of 16 computer units vec5, Cayman has 24 blocks of 16 computer units vec4. We are dealing with 320 units against 384 units vec5 vec4, which is less flattering when incorporated into cores because it gives us 1536 cores only cons for Cayman 1600 for Cypress. An important detail, however, is found at the texturing units whose number is fixed at 4 per block. Cayman to see his power at this level increased from 20% to equal frequency. Note that AMD has reported increasing flow calculation in double precision, but it is a twisted way of interpreting the fact that the "one" does not support double precision. A unit of Cayman is identical to a unit of Cypress at this level.

AMD did not stop there and introduced other minor improvements to its architecture. The first concerns the treatment of geometry who is parallelized in order to break the limit of a triangle per cycle. Nvidia retains an advantage with small triangles and above, with more units of simple geometric processing, avoiding a bottleneck at the GPU when extensive data are generated by the tessellation. To combat this problem, AMD has expanded the buffer dedicated to Barts, the GPU Radeon HD 6800, and Cayman and goes further with it which is capable of transferring all this data temporarily in video memory to avoid blocking while the GPU . This function is not directly exposed and we do not know if it engages automatically when certain charges or if AMD is to be used manually on a case by case basis.

AMD has also improved ROPs to increase throughput formats 16-bit integers and 32 bit float. They are also becoming more efficient with the antialiasing, as well as during the memory write mode compute. In this regard, AMD was inspired by what Nvidia has and allows the simultaneous execution of several different kernels, whereas before the GPU would assign successive periods of execution. It's the same for communication with the CPU that can be done in both directions simultaneously with 2 DMA engines, as GF100/110. The memory controllers have also been revised to more easily withstand the fast GDDR5.

[Charlton]

Specification :

AMD Mobility Radeon HD 6970M Crossfire (or ATI Mobility Radeon HD 6970 X2) shall consist of two HD 6970M in Crossfire composite (similar to SLI). The two graphics cards to render normally alternately a picture. This can also come to Microruckler (non-uniform distances between two images, despite stuttering lead to significant liquid fps rate of about 30 fps). Since the memory chips in each graphics card has the same content are stored, you can not add the amount of memory. 2x1 GB are thus as 1 GB for games. The Radeon HD 6970M is based on the desktop HD 6850 graphics chip, code-named Blackcomb XT. The Blackcomb XT supports UVD3 chip to decode HD videos and Eyefinity + for connection of up to 6 monitors (when the notebook offers DisplayPort).

With a frequency slightly higher GPU, the Radeon HD 6970 has an output of calculation almost identical to that of the Radeon HD 5870. It benefits from a flow cons of triangles and filtering more important. Good news, the Radeon HD 6950 is less castrated compared to the Radeon HD 6970 than was the Radeon HD 5850 compared to the Radeon HD 5870.

Consumption

Like any good high end card that follows, the Radeon HD 6970 is impressive. It measures 27.5 cm long, opts for a cooling system completely standard and feed taken in watts via two PCI-Express (6 and 8 pin). For our tests of consumption, we pushed the slide of the Power Control to find the most effective PowerTune. As a reminder, this new feature allows the card to assay the following consumer needs of the chip. As for the GeForce GTX 570, we realized that our test program was listed by the bullets of the card. This flange is thus itself and it has never been possible to exceed 225 watts of consumption (as against 250 announced). If we look closer, the graphics chip was not pushed to its maximum capacity (because of PowerTune). Under these test conditions, the chip does not exceed 85 ° C and the fan runs at 40% of its maximum speed (2265 rpm). To the ear, it is a bit louder than the GTX 570. It is therefore very difficult to pronounce a firm and final on the maximum consumption of the HD 6970 and the ability to warm up to the computing unit.

Next games and DirectX version used, the gaps are widening between the GTX 570 and HD 6970. Evacuate immediately Crysis Warhead, where scores of both cards are neck and neck, to focus on more meaningful and representative titles of graphics engines currently used by development studios. Dirt 2 under DirectX 11, the GeForce offers 10 to 15 frames per second (fps) more than the new AMD. By cons, in S.T.A.L.K.E.R.: Call of Pripyat the gap between the two chips is between 4 and 6 fps, always in favor of Nvidia. DirectX 10, whether in Resident Evil 5, Far Cry 2 or HAWX, first name, Nvidia is definitely leading the discussion, but the Radeon is doing very well also, and consistently approaches or exceeds 100 fps Full HD. A quick glance at the scores in the native resolution of our 30-inch screen shows us that the cards go pixel to pixel and that variations in frames per second are not very revealing.

[Charlton]

CrossFire

These two new Radeon share as the cooling system as PCB and power stage. Small differences are visible at his level but it is only details of editing and production, both PCBs do not visibly out of the factory. It is possible that the Radeon HD 6950 trade, especially once the first lots sold were slightly modified in an obvious manner since they do not need a power stage too sturdy. She also just 2 6 pin power connectors 8 +6 cons for the Radeon HD 6970. In addition to using the very fast GDDR5, Hynix brand certified to 6 Gbps or 1.5 GHz to 3 GHz and its management for sending data, AMD has decided to install a total of 2 GB by default on the Radeon HD 6900. Otherwise it is an AMD design classic with a turbine and a steam room. Like the Radeon HD 5870 but unlike the Radeon HD 5850, an aluminum plate covers the back of the PCB. In terms of connectivity, AMD has taken the one introduced with the Radeon HD 6800, i.e. 2 Dual-Link DVI, HDMI outputs and 2 mini-DisplayPort. Sorting and Quad

CrossFire X is supported via two dedicated connectors. Besides these, AMD has introduced a small switch that can trigger a backup bios. The first bios can be flashed by the user, while the latter is protected, thereby avoiding any problems during an upgrade.

The maximum allowable consumption level is fixed at 250W for the Radeon HD 6970 and 200W for the Radeon HD 6950. If the GPU is estimated that consumption of the card goes beyond that, it gradually reduces its frequency technology via PowerTune. AMD, however, have fixed the precise boundaries of these maps so that such protection does not engage in the most demanding games and in the worst situations, that is to say with a copy of the GPU to leakage currents high placed in a hot environment. You can change this limit consumption via the Catalyst Overdrive panel. AMD can be adjusted between -20 and +20%, between 200 and 300W for the Radeon HD 6970. The reduction can reduce consumption and pollution in games the most demanding situations while increasing it allows you to have more headroom for overclocking, otherwise an increase in frequency could change nothing if the GPU is limited to level.

Radeon HD 5870

The Radeon HD 5870 may retire, succession is almost assured. On average, the Radeon HD 6970 is superior to 20%. A good number, but still pretty average when you puts the card into the premium segment. We are not persuaded that the adjustments made on the chip architecture are quite relevant as it is and it feels like for some games in which the earnings are less than 8% of a chip to another Radeon the same definitions. AMD has once again on the price to convince the Radeon HD 6970. The user is certainly not a bad deal, but given our test results a bit disappointing, enticing specifications of the HD 6970 are misleading. Especially as on top models, users are generally a little less stingy. In our view, the GeForce GTX 570 is a better investment over the long term and, as expected, the GeForce GTX 580 remains sovereign over the high end.

Conclusion:

Obviously, the pair of AMD 6970 in CrossFire X has a yield of over 6970 alone. Is obvious. However, in our tests has been only a 15% improvement in games, and 25% in synthetic benchmarks, which in truth is quite low. Frankly, we expected more. AMD promised us in the presentation of its 6970 and 6950 that had improved the efficiency in CrossFire X configurations over previous generations. The results speak for themselves, and while not reaching double the performance (which is impossible for certain reasons at low level) if it were desirable to achieve as low a yield of between 30 and 40% above the configuration of a graphic.

To achieve that just 15% have to sacrifice a huge energy consumption, nearly twice that in the configuration of a graph. To achieve maximum power of 625 watts, a real lot to gain only a few frames per second. Compared to the temperatures we must recognize that rise but remain at levels successful, below 90 degrees Celsius. For its part, noise is not excessive, although easily detectable. After comparison of CrossFire X opposite the configuration of a single graphics card, we turn to face it on other models in the market, and obviously we look at direct rival: NVidia GTX 580.

The price of these two models are weighted in favor of NVidia. For its part, the yield is higher in the 6970 CrossFireX but only 10% in games and 25% in benchmarks, even with very similar temperatures but energy consumption almost double. With data in hand we must conclude that AMD 6970 is fine, but the CrossFire X needs improvement.

For twice the price get higher performance, but in a small measure. It is important to note that NVidia has a graphical performance 2xAMD quite similar to 6970 but in a single graphics card and set a price substantially. At this point, is worth the SLI / CrossFire X? Interestingly, there have been some models as the nVidia GTX 460 have been fantastic in this respect.