NVIDIA GeForce GTX 470M SLI vs NVIDIA GeForce GTX 485M
NVIDIA GeForce GTX 470M SLI► remove
The NVIDIA GeForce GTX 470M SLI is a high-end graphics solution for laptops based on two GeForce GTX 470M graphic cards in SLI mode. Each card usually renders a single frame in an alternating manner. Therefore, it may suffer from micro stuttering in low fps ranges of 30fps. This happens because of different timespans between two frames (e.g. Irregular delays between sequential frames). Each single GTX 470M is based on the GF104 core and therefore supports DirectX1 11 and OpenGL 4.0.
The GF104 core of the GTX 470M is related to the GF100 core of the GeFore GTX 480M that offers 384 shaders and a 256 Bit memory bus for GDDR5. The GTX 470M, however, offers only 288 cores of the 384 and a 192 Bit memory bus. The architecture of the GF104 is not comparable to the old GT215 (e.g., GeForce GTS 350M) or G92b (e.g., GeForce GTX 285M) cores.
Unlike the GF100, the smaller GF104, GF106, and GF108 cores were shortened and considerably modified. In contrast to the GF100, which was designed for professional applications, these chips target the consumer market. They feature more shaders (3x16 instead of 2x16), more texture units (8 instead of 4) and SFUs per streaming multi-processor (SM). As there are still only 2 warp schedulers (versus 3 shader groups), Nvidia now uses superscalar execution to use the higher amount of shaders per SM more efficiently. In theory, the shaders can therefore be utilized more efficiently and the performance per core is improved.
However, in worst case scenarios the performance can also be worse than that of the GF100 (and its predecessors). The ECC memory protection, which is important for professional applications, was completely omitted and the FP64 hardware was shortened (only 1/3 of the shader are FP64-capable and therefore only 1/12 of the performance of the FP32). Because of these cutbacks, the size of the SM grew only by 25% despite the higher number of shaders and larger warp schedulers with superscalar dispatch capabilities. Due to the different shader architectures and the higher clock rate of the shader domain, the core count can not be directly compared to AMD cores of the Radeon 5000 series (e.g. HD 5850).
Detailed information on the GF104 architecture (and by extension also the GF106 and GF108) can be found in the desktop GTX 460 article by Anandtech.
The performance of the GTX 470M SLI combination should be similar to the GTX 480M SLI due to the high-clocked GTX470M cards. Generally, the SLI gain depends on the used application/game and the driver support. Currently, the SLI driver support from Nvidia is better than the Crossfire support from AMD, leading to higher gains. About 30% more performance is gained on average compared to a single GTX 470M (in high resolutions and with Anti-Aliasing). Therefore, all games from 2010 except Metro 2033 should be playable in the highest detail settings. Metro 2033 only ran with 23 fps in our tests with a GTX 480M SLI system and should therefore score similar on the GTX 470M SLI combination.
A novelty of the GF104/106/108 chips is the support of Bitstream HD Audio (Blu-Ray) output via HDMI. Similar to the Radeon HD 5850, the GTX 470M can transfer Dolby True HD and DTS-HD bitstream without quality loss to a HiFi receiver.
The GTX470M offers the PureVideo HD technology for video decoding. The included Video Processor 4 (VP4) supports feature set C and therefore the GPU is able to fully decode MPEG-1, MPEG-2, MPEG-4 Part 2 (MPEG-4 ASP - e.g., DivX or Xvid), VC-1/WMV9, and H.264 (VLD, IDCT, Motion Compensation, and Deblocking). The X500 tester was able to decode the VC-1 encoded Elephants Dream video with about 3-6% CPU load (according to the task manager). The H.264-encoded Big Buck Bunny video was played back with 1-3% CPU load (both 1080p videos).
Furthermore, the GPU is able to decode two 1080p streams simultaneously (e.g. for Blu-Ray Picture-in-Picture).
Through CUDA, OpenCL and DirectCompute 2.1 support, the GeForce GTX 470M can assist in general calculations. For example, the stream processor can encode videos considerably faster than certain high-end CPUs. Furthermore, physics calculations can be done by the GPU using PhysX (e.g. supported by Mafia 2 or Metro 2033). SLI systems can also dedicate a single card for PhysX calculations.
According to Nvidia, support for 3D Vision on the GTX graphics cards is also new. It enables the laptop to send 3D contents (3D games, 3D Web Streaming, 3D photos, 3D Blu-Rays) to a built-in 3D enabled screen or an external 3D TV (only if supported by the laptop manufacturer).
Unofficially, the power consumption of the GeForce GTX 470M should be about 75 Watt (TDP including the MXM board and memory), which is about the level of the Mobility Radeon HD 5870. Therefore, the SLI system needs 2x75 Watt and therefore a lot less than the GTX 480M SLI with 2x100W. Without load, the chip is clocked at 50/100 MHz (chip/shader) in 2D respectively 200/400 in 3D mode to save power.
Compared to desktop graphics cards, the GTX 470M SLI combination is most similar in performance to a GTS 450 SLI combo, as the desktop GTX 470 is significantly faster.
NVIDIA GeForce GTX 485M► remove
The Nvidia GeForce GTX 485M is the fastest graphics card for laptops at the time of announcement (Q1 2011). It is based on the GF104 chip and offers all 384 shader cores and the full 256 Bit memory bus. Due to the high clock rate of 575 MHz, it is significantly faster than the old GeForce GTX 480M of which it replaces. It also supports DirectX 11 and OpenGL 4.0.
Other than the GeForce GTX 480M, the 485M is no longer based on a trimmed down GF100 chip, but on the related GF104 instead. The GF104 is designed for the consumer sector and has a total of 384 cores. A number of cores may be disabled, for example the 470M with only 288 active cores.
The technology of the GF104 differs quite a bit from the GF100 chip (which was actually designed for professional use). The GF104 has more shaders (3x16 vs. 2x16), texture units (8 vs. 4) and SFUs (Special-Funciton-Units) per Streaming-Multiprocessors (SM). Nvidia now uses the superscalar architecture as there are still only two warp schedulers supporting three shader blocks. In theory, this helps to utilize the shaders more efficiently and increases the performance per core.
However, in the worst case, the performance can drop below the GF100 architecture (and its predecessors). The ECC memory protection, important in professional applications, was completely omitted and the FP64 was trimmed down (only 1/3 of the shaders are FP64-capable, only 1/12 of the FP32 performance). Because of these reductions in the GF104, the size of a SM increased only by 25% despite the higher number of shaders.
Note that it is not possible to directly compare the number of cores to the AMD Radeon graphics cards (e.g. HD 5870) or even to Nvidia's own predecessors (e.g., G92b), because shader architecture and clock rates are significantly different in the GF104 chip.
In our extensive test of the GeForce GTX 485M, we found that the GTX485M is significantly faster than the old GeForce GTX 480M (at the same TDP rating). The performance is on a level with two GeForce GTX 460M in SLI mode. Nearly all games are therefore playable in highest details and resolutions. Even demanding games like Mafia 2 or Battlefield Bad Company 2 can run fluently in 1080p with maximus detail settings. Detailed benchmarks can be found at the end of this page.
What's new compared to the GF100 is support for Bitstream transfer of HD Audio (Blu-Ray) via HDMI in the GF104 chips. Similar to the HD 5850, the GTX 485M can transmit Dolby True HD and DTS-HD via Bitstream to compatible receivers without quality loss.
For decoding HD videos, the GTX485M supports PureVideo HD. The built-in Processor 4 (VP4) handles Feature Set C. As a result, MPEG-1, MPEG-2, MPEG-4 Part 2 (MPEG-4 ASP - z.B. DivX or Xvid), VC-1/WMV9 and H.264 can be fully decoded by the graphics card (VLD, IDCT, Motion Compensation, and Deblocking). Furthermore, two streams can be simultaneously decoded in realtime, e.g. Blu-Ray Picture-in-Picture (2x1080p lt DXVAChecker). In addition, PureVideo HD indicates HDCP encoding for digital interfaces.
The shader cores (also called CUDA cores) can also be used for general computations (e.g. Video Transcoding) by using the interfaces CUDA, DirectCompute 2.1 or OpenCL. Thanks to PhysX, the 485M can also perform physics calculations.
According to Nvidia, the support for 3D Vision includes support for the recent HDMI 1.4a standard as well. If enabled by the laptop manufacturer, content such as 3D games, 3D web streaming videos, 3D pictures and 3D Blu-Ray videos can be displayed on a 3D-capable TV (via discrete 3DTV Play) or on the internal notebook 3D display.
With regards to energy demand, the GTX 485M should be on par with the GeForce GTX 480M. In other words, both graphics cards should draw about 100 Watts each when including their respective memory and MXM boards. Due to the higher performance from the 485M, the performance/power efficiency here has clearly been improved.
Compared to desktop graphics cards, the performance of the 485M should be on par with a GeForce GTX 460 768MB which features less cores but operates on higher clock rates.
|NVIDIA GeForce GTX 470M SLI||NVIDIA GeForce GTX 485M|
|GeForce GTX 400M Series|
|Cores||576 - unified||384 - unified|
|Core||535 MHz||575 MHz|
|Shader||1070 MHz||1150 MHz|
|Memory||1250 MHz||1500 MHz|
|Bus||192 Bit||256 Bit|
|DirectX||DirectX 11, 5.0||DirectX 11, 5.0|
|Technology||40 nm||40 nm|