Radeon HD 6000 Series

The Radeon HD 6000 series is a series of desktop graphics chips from AMD, which was introduced in the fall of 2010 as the successor to the Radeon HD 5000 series. All graphics processors of this series support Shader Model 5.0 DirectX 11 and OpenGL 4.1 after. The Radeon HD 6000 series, which also carries the code name "Northern Iceland " is, for the first time no longer marketed under the logo of ATI, but directly under AMD's logo. The sequel is called " AMD Radeon HD 7000 series " and came in spring 2012 on the market.

  • 5.1 Measurement of power consumption of a whole system

Description

The first models of the Radeon HD 6000 series were presented in form of the Radeon HD 6850 and 6870 on 19 October 2010, the trade press, before three days later officially dropped the NDA. Both cards use while the Barts GPU, which continues on the " Terascale 2 architecture" of the Radeon HD 5000 series is based and its shader units are thus constructed in 5D VLIW design. Also identical to the cluster structure of the GPU (16 shader and four texture units ), which is beard with a total of 14 clusters from 224 shader 56 and texture units. Since the AMD Barts GPU does not map to the high-end sector, they waived the double-precision support, which transistors could be saved. Overall Barts consists of approximately 1.7 billion transistors on a die area of 255 mm ². After Nvidia with its Fermi architecture could have advantages in the field of Tessellationsleistung, AMD revised the Tesselator for the Barts GPU. Thus, the "Fixed Function Unit", the thread management and the data buffers have been improved, which claims to performance over the Cypress GPU has doubled with moderate Tessellationeinsatz AMD. However, the advantage does not apply to larger Tessellationbelastungen again. Besides the possibility for 3D playback, a " Unified Video Decoder" third generation is implemented, which supports hardware acceleration for more video formats.

The Radeon HD 6870 uses a uncircumcised Barts GPU, achieving a performance that lies between the Radeon HD 5850 and the GeForce GTX 470. On the Radeon HD 6850 two of the 14 shader clusters are disabled, so they roughly the power of the GeForce GTX 460 ( 1024) achieved, which was until now been relatively successful in the performance sector. In some test reports, the Radeon HD 6850 was initially " too positive " rating, as the test samples partially all shader clusters were unlocked and corresponding map thus reached a power which is not that corresponded to the commercially available cards. In the trade press sometimes controversial were the naming and the AF filter: Thus, although a native of the Evergreen series banding effect ( abrupt texture transitions) has largely been resolved, but the AF filter also tended more to flicker, which among other things due to a lowering of the default filter quality in the Catalyst driver is what had been done also on the older models of the Radeon HD 5000 series with Cypress GPU. Eventually took AMD Catalyst drivers and 11.1a 11.2 lowering the default filter quality back, whereby the fact of the flickering filter AF could be solved at least partially. The name made ​​so far for confusion, as the Radeon HD 6850 and 6870 compared to the models Radeon HD 5850 and 5870 showed a lower performance, although the name suggests the opposite.

On 15 December 2010, AMD Radeon HD 6950 and the 6970 before. These are based on the Cayman GPU, for which a new shader design was developed. A shader architecture 5D VLIW arrangement has been used for the R600 GPU, were combined with the five- stream processors (ALU ) to a shader. There are four ALUs were responsible for the calculation of the RGB and the alpha value, the fifth acted as a T -Unit for Special Function functions such as sine and cosine calculations. The disadvantage of this arrangement is that if in one clock cycle a value is not needed, the appropriate stream processor then lying idle. According to AMD, the utilization was this arrangement, at around 70 %. In the Cayman GPU was changed this arrangement in order to favor a 4D VLIW system. This related to the T -Unit, for their tasks now three ALUs are connected together. This is to improve the utilization and have been reduced the number of required transistors. Another advantage of the new arrangement is that instead of five, only four ALUs is expected to double precision needed. Thus, the DP value is now 25% of the SP value, instead of 20% previously. After Nvidia could have distinct advantages in Tessellationperformance with the Fermi architecture, AMD introduced the tessellator on the Cayman GPU from twice. Ultimately, the Cayman GPU of around 2.64 billion transistors on a die area of 389 mm ² and is home here in 24 shader clusters 1536 stream processors, 96 texture units and 32 raster.

The introduction of the Radeon HD 6950 and 6970 was delayed initially, with AMD did not specify the reasons. The Radeon HD 6970 roughly pointed to the performance of the Geforce GTX 570. As had been suspected in advance because of the architectural changes that the Radeon HD 6970 would compete with the GTX 580, they could not meet the sometimes very high expectations. This also contributed to the already known from the Barts GPU Flimmerneigung the AF filter unchanged also appeared on the Radeon HD 6970. Also, the power consumption was, despite new PowerTune option, criticized, although this was not higher than in the same fast Geforce GTX 570 However, the Radeon cards of the previous generation often reported with the same performance at lower power consumption than its Nvidia competitor. Benefits was the Radeon HD 6970 show at very high resolutions, partly due to the large 2 GB of video memory.

In the trade press significantly more positively than the Radeon HD 6970 was the 6950, the Radeon HD 5870 to replace. This was due, inter alia, the fact that Nvidia had no competition first card is available, but also that they yielded only about 10% less power than the Radeon HD 6970, but it consumed up to 30% less power. On the 6950 two shader clusters of the Cayman GPU were disabled, so even 1408 stream processors were available. However, it was possible to reactivate with cards of the first batches with a so-called Bios flash the disabled cluster, as these had not yet been processed with a laser cut. This disconnection has been favored by the new DualBIOS, but led to loss of warranty. While a number of board partners independently created variants of the Radeon HD 6950 with 1GB of memory, AMD presented these officially on January 25, 2011. Bonds were placed by AMD against the same day they launched Geforce GTX 560 Ti.

On 8 March 2011 the last model was followed based on the Cayman GPU, the Radeon HD 6990th On this two Cayman GPUs were placed in full extension. This dual- GPU graphics card turned out in their presentation first as the fastest card on the market. However, the map also set up new negative records in the power consumption and noise. AMD built in 6990 two 8 -pin power connector, the TDP thus increased to 375 watts and thus broke the official PCI -SIG specifications. At the same time AMD equipped the card with a second bios, which raises the core clock speed of 830 to 880 MHz in order to achieve the clock level of the Radeon HD 6970. With the use of alternative Bios, AMD as " AUSUM " means ( Antilles Unlocking Switch for Uber Mode), but also lost the warranty, and the TDP rises again to 450 watts ( power consumption, however, can be up to 506 watts in). Overall, the product presentation looked more like a feasibility study, because with the increased consumption also increases the demands on the cooling system. These requirements, however, you will not meet with the built-in cooler. The volume increases as compared to other products enormously. With the alternative Bios " AUSUM " it is (some with over 14 Sone ) unreasonable due to the continuing high noise to operate the computer in the same room in which you play as well.

4 February 2011, AMD Radeon HD before the models 6450, 6570 and 6670. The measures provided for the low -cost sector maps were initially only intended for the OEM market, the Radeon HD 6450 is based on the Caicos GPU. The Radeon HD 6570 and 6670, however, the use of 716 Mil. Transistors existing Turks GPU. Already on January 20, AMD had the cards Radeon HD 6750 and 6770 presented, which is 5750 and 5770 is reprints of models Radeon HD who were also intended for the OEM market. On 7 April, the AMD Radeon HD 6450 changes also to the retail market before, with the actual launch only took place on April 19. The models Radeon HD 6570 and 6670 were released on this day for the retail market.

GPUs

Naming

In the Northern - Iceland - series the same designation system, as they have for the Radeon HD 3000 series control panels. All graphics cards are referred to as "AMD Radeon HD" and an additional four-digit number that generally begins with a "6" ( for the series ). The second and third digits are the subdivision into different models.

Distribution:

  • HD 64xx: low-end
  • HD 65xx, 66xx and 67xx HD: Mainstream
  • HD 68xx: Performance
  • HD 69xx: High- End

Given the general slump in market prices and currency fluctuations, the original classifications of AMD are not applicable in principle.

Model data

Notes:

  • The specified clock rates are recommended by AMD or specified, the memory clock of the I / O clock is specified. However, may differ by several clock a few megahertz of the exact timing, further is the final determination of the clock rates in the hands of the respective graphics cards manufacturers. Therefore, it is quite possible that there are graphics cards models or will be having different clock speeds.
  • With the stated date of the appointment of the public presentation is given, not the date of availability of the models.

Performance

For each model, the following theoretical power data:

Notes:

  • The above performance data for the computing power of the stream processors, the pixel fill rate, the Texelfüllrate and memory bandwidth are the theoretical maximum values. The overall performance of a graphics card depends, among other things, on how well the available resources can be exploited or utilized. In addition, there are other not listed here, factors that affect the performance.
  • The computing power of the stream processors is not directly comparable with the performance of Nvidia Geforce series, since this is based on a different architecture, which scales differently.

Power consumption data

The measured values ​​listed in the table refer to the pure power of graphics cards that comply with the AMD reference design. In order to measure these values ​​, it requires a special measuring apparatus; depending on the employed measurement technique and given measurement conditions, including the shared program, with which the 3D load is generated, the values ​​between different equipment may vary. Therefore, here measured value ranges are specified, each representing the lowest and highest measured values ​​from different sources.

Far more common than the measurement of the consumption of the graphics card is to determine the power consumption of a whole system. To this end, a reference system is assembled, in which the various graphics cards to be installed; then finds the measurement using an energy cost meter or similar apparatus rather than directly to the outlet. However, the significance of the measurements is limited: it is not clear which consumption comes from the graphics card and what is attributable to the rest of the PC system. The difference in consumption between idle and load 3D operation of this measuring method is not only dependent on the program with which the load was produced; the utilization and efficiency of the rest of the PC system, including the power supply, motherboard and processor affect the measured difference also. Because different the tested systems are usually of one's own home PC system, the values ​​given there can not be mapped to its own system. Only data from otherwise identical systems are good (conditional) for comparison with each other. Because of this dependence overall system measurement values ​​are not listed in the local table. But since they can provide a better picture of the practical power of a concrete system with a certain graphics card, web pages are listed under the links that are undertaking such measurements.

56259
de