by Anton Shilov
11/29/2011 | 09:52 PM
In an attempt to offer an ultimate graphics solution for the holiday season, Nvidia Corp. introduced this week a new enhanced GeForce GTX 560 Titanium 448 core Limited Edition that embraces GF110 graphics processor and comes with 448 stream processing units. The graphics board gives more horsepower than the regular GeForce GTX 560 Ti (with GF114 chip) and will allow Nvidia to get rid of excessive GF110 chips ahead of next-gen launch.
Nvidia GeForce GTX 560 Ti 448 Cores Limited Edition is based on the cut-down GF110 graphics processing unit (GPU) with reduced amount of stream processors (SPs), texture units (TUs), render back ends (RBEs) and memory channels. The GeForce GTX 560 Ti-448 LE has 448 SPs, 56 TUs, 40 RBEs and 320-bit memory bus. The base clock of the chip is 732MHz, stream processor operates at 1464MHz and memory functions at 3.80GHz. Nvidia recommends its partners to install 1280MB of memory onto the new graphics cards.
The new GeForce GTX 560 Ti-448 LE offers slightly higher (1.311TFLOPS) compute power compared to the original GeForce GTX 560 Ti (1.26TFLOPS), tangibly higher memory bandwidth, significantly higher tessellation performance as well as 3-way multi-GPU SLI support. But the novelty comes at higher price (it costs $289), consumes more power (210W vs. 170W) and is longer (10.5" vs. 9") than its brother, which will not allow it to be installed into all computer cases.
Nvidia clearly aims GeForce GTX 560 Ti-448 LE at enthusiasts who did not get GeForce GTX 570 earlier and are currently looking forward a graphics card upgrade ahead of the holiday season. Thanks to the lack of any well-positioned competitors from AMD, the model GTX 560 Ti-448 LE may become a viable choice for many gamers this year.
Just like the GeForce GTX 560 Ti OEM launched earlier this year, the GF110-based GeForce GTX 560 Ti-448 LE to get rid of chips that have faulty stream processors and cannot power GeForce GTX 580 (512 SPs) and GeForce GTX 570 (480 SPs). The approach has proved to be very efficient on previous-generation processors.