San Jose (CA) - PCI Express allows Nvidia to tap into system memory to save expensive on-board graphics memory and achieve high performance at the same time. Tom's Hardware Guide and other publications however raised concerns over a possibly misleading naming of the cards and prompted Nvidia to revise the product name.
Using system memory has been an idea of graphics card manufacturers in the late nineties to be able to offer their products at cheaper price points. Now the approach returns with ATI's recently announced "Hypermemory" and Nvidia's "Turbocache". This time, however, the firms are able to use a much faster interface to write and read data from the mainboard: PCI Express supplies four Gbit of bandwidth between the core logic and the graphics processor.
As it was the case with the "shared memory" from the late nineties, Nvidia intended to highlight the actual memory capability of the cards which referred to the on-board memory in addition to the memory used from the mainboard. Originally, the cards were planned to hit the market as "GeForce 6200 w/ TurboCache supporting 128 MB @ 10.8 GB/s" and as "GeForce 6200 w/ TurboCache supporting 128 MB @ 13.6 GB/s".
After concerns raised by Tom's Hardware Guide as well as "other publications" that the designations could be confusing or misleading, Nvidia informed us of a last minute name change: Now the cards will carry the names "GeForce 6200 TurboCache 16-TC/128 MB" and "GeForce 6200 TurboCache 32-TC/128 MB", respectively. There is also a version with 64 MByte, called the "GeForce 6200 TurboCache 64-TC/256 MB".
According to Nvidia spokesman Brian Burke, it is "very unlikely" that stores will receive graphics cards carrying the original product naming. Suggested retail pricing for the cards starts at $80 for the 16-TC/128 MB and tops out at $130 for the 64-TC/256 MB.
ATI takes another look at shared memory