Powered by Blogger.

Nvidia GTX 760 vs AMD Radeon HD 8000

AMD and Nvidia never have time to rest on their laurels, instead playing a constant game of leapfrog where the biggest winners are gamers.

NVIDIA GeForce GTX 760 Specifications

The GeForce GTX 760 is equipped with GK104 processor, a cut-down version with 1152 CUDA cores on board. As I told you earlier, this card was not made as an upgrade to GeForce 600 series, it’s rather an offer for GTX 560 Fermi users (providing twice the performance). The GTX 760 will be replacing the GTX 660 Ti in the current line-up. Thus, you should not be surprised that the GTX 760 will use the reference design known from GTX 670, 660 Ti, 660 and 650 Ti.

The GeForce GTX 760 is clocked at stock 980 MHz for base and 1033 MHz for boost, whereas the boost state will depend on many factors. The GTX 760 supports GPU Boost 2.0, which is temperature-driven technology, so the cooler the card, the higher the clock you’ll get. As it turns out, the 2GB GDDR5 memory will be a default configuration for the GTX 760, but 4GB models will also be offered by AIB partners. The memory clock was set to 1502 MHz, which gives 6 GHz effective speed. The GTX 660 Ti replacement is equipped with a wider 256-bit interface.

The new mid-range model will require a pair of 6-pin power connectors, consuming up to 170W in typical operation. Also the standard set of display outputs will be offered with the card: DVI-I, DVI-D, HDMI and DisplayPort. Card will be officially announced on June 25th. The pricing was not yet revealed.
-->
AMD Radeon HD 8000 series

On the other hand, AMD also released its flagship product the Radeon HD 8000 series. Chiphell posted some updated specifications for AMD's new crop of cards. The fastest of the bunch is the Radeon HD 8970, which features 36 compute units, 2,304 stream processors, 144 TMUs, 48 ROPs, and an 1,100MHz core clockspeed. It also has 6GB of GDDR5 memory clocked at 7,000MHz (effective) on a 384-bit bus, all of which results in 5.07TFLOPS of floating point performance. Versus the Radeon HD 7970 GHz Edition, the new part will run 35 percent faster and cost $599.

Next up is the Radeon HD 8950 ($399) with 28 compute units, 1,792 stream processors, 112 TMUs, 32 ROPs, 4GB of GDDR5 memory clocked at 7,000MHz on a 256-bit bus, 1,200MHz core clockspeed, and 4,30TFLOPS of floating point performance. It's said to run 10 percent faster than a Radeon HD 7970 GHz Edition card.

Below that sits the Radeon HD 8870 with 24 compute units, 1,536 stream processors, 96 TMUs, 32 ROPs, 2GB of GDDR5 memory clocked at 6,000MHz on a 256-bit bus, 1,100MHz core clockspeed, and 3.38TFLOPS of floating point performance. This $299 card is supposed to be 40 percent faster than a Radeon HD 7870.

Finally there's the Radeon HD 8850 ($229) with 20 compute units, 1,280 stream processors, 80 TMUs, 32 ROPs, 2GB of GDDR5 memory clocked at 6,000MHz on a 256-bit bus, 1,000MHz core clockspeed, and 2.56TFLOPS of floating point performance. It's said to run 15 percent faster than a Radeon HD 7870 graphics card.

The second half of 2013 only gets more and more heated up with excitement...!.