NVIDIA GeForce 6200 with TurboCache
TurboCache Explored
One of the main reasons NVIDIA is launching the GeForce 6200 with TurboCache technology, is to bring the same feature set found in their flagship GeForce 6800 Ultra down to a much more affordable price point. This chart shows where the GeForce 6800 series, 6600 series and 6200 series differ.
As you can see, the only feature not found on the 6200 is SLI (Scalable Link Interface). SLI is NVIDIA's proprietary technology for using multiple PCI Express graphics cards in tandem to split 3D graphics workload and increase performance dramtically. That feature is reserved for only NVIDIA's high-end gaming cards. Other than SLI though, the GeForce 6200 has all of the same capabilities as its big brothers. NVIDIA learned with the GeForce 4 MX, that it does't make sense to release an entry level video card that doesn't have same features of the higher-end cards from the same generation. That's why last year we had the GeForce FX 5200, and now the GeForce 6200. For a more in-depth explanation of each of these technologies, we recommended checking out our initial look at the NV40 here, our coverage of the GeForce 6600 and it's updated video engine, and our recent look at NVIDIA's SLI technology here, and here.
 |
|
 Traditional 3D Pipeline |
 GeForce 6200 with TurboCache Pipeline |
To implement their TurboCache technology, NVIDIA had to modify the GeForce 6200's pipeline somewhat to incorporate a new memory management unit. The MMU allows the GeForce 6200 with TurboCache to seamlessly allocate and de-allocate surfaces in system memory, as well as read and write to that memory efficiently. Additionally, architectural modifications were made to few of the pipeline elements to handle the increased latency that occurs from accessing system memory across the PCI Express interface (sections highlighted in yellow are the ones that have been modified). As you can see by the block-diagram, the GeForce 6200 with TurboCache is a 4-pixel pipeline GPU, with three vertex shaders.
;)
The hardware wasn't the only thing that had to be modified to enable 'TurboCache' though. NVIDIA's Forceware drivers were also modified to so they could determine where color, texture, and z-buffer data is located. This new feature incorporated into the driver helps to maximize performance for each application. The TurboCache Manager (TCM, as it is known), lets additional memory be allocated for graphics usage based on a particular application's needs. And then once that application is closed, the memory that was allocated for graphics is released and made available for system use. This entire process is done behind the scenes and and is transparent to the end-user. As you can see in the screenshot above, the driver actually reports 128MB of available memory, even though a card with only 16MB or 32MB of local memory is installed in the system.
