GeForce GTX 680 - NVIDIA's 28nm Successor to Fermi

The NVIDIA GeForce GTX 680

NVIDIA GeForce GTX 680 - Stacking Up Against the Competitors

The reference GeForce GTX 680 card looks vaguely similar to the GeForce GTX 580. Its differences lies skin-deep with its 1536 CUDA cores and eight Polymorph Engines (version 2.0). GeForce GTX 680 also features a GPU fan made from an acoustic dampening material, reducing the noise it emits as it spins to cool the card. The second feature is its heatsink with an improved finstack for better airflow to ensure optimized cooling.The third feature are the triple heatpipes that resemble a pretzel in direct contact with the Kepler GPU for better heat transfer to the heatsink.

Apart from making the GeForce GTX 680 fast and superior to its older series, one of its design goals was to make it the quietest ever top-end graphics card and we believe they succeeded with all the efforts put into the cooler design. For a closed chamber type of cooler, the GeForce GTX 680 was miles ahead of the competition in being quiet, no matter if it was on idle or load throughout our testing,

The finstack of the heatsink that sits beside the GPU fan made from acoustic dampening material. If NVIDIA can pull out such stops for its reference card, we wonder what the card manufacturers would have to do to go above this bar that NVIDIA has set with these features.

GeForce GTX 680 and competitive SKUs compared

Model NVIDIA GeForce GTX 680

AMD Radeon HD 7970

AMD Radeon HD 7950 NVIDIA GeForce GTX 580 AMD Radeon HD 6970
Core Code GK104  Tahiti XT  Tahiti Pro GF110 Cayman XT
Transistor Count 3540 million  4300 million  4300 million 3000 million 2640 million
Manufacturing Process 28nm 28nm 28nm 40nm 40nm
Core Clock 1006MHz 925MHz 800MHz 772MHz 880MHz
Stream Processors 1536 Stream processing units  2048 Stream processing units 1792 Stream processing units 512 Stream processing units 1536 Stream processing units
Stream Processor Clock 1006MHz  925MHz 800MHz 1544MHz 880MHz
Texture Mapping Units (TMU) or Texture Filtering (TF) units 128  128 112 64 96
Raster Operator units (ROP) 32  32 32 48 32
Memory Clock 6008MHz GDDR5  5500MHz GDDR5 5000MHz GDDR5 4000MHz GDDR5 5500MHz GDDR5
DDR Memory Bus 256-bit  384-bit 384-bit 384-bit 256-bit
Memory Bandwidth 192.3GB/s  264GB/s 240GB/s 192.4GB/s 176GB/s
PCI Express Interface PCIe ver 3.0 x16  PCIe ver 3.0 x16 PCIe ver 3.0 x16 PCIe ver 2.0 x16 PCIe ver 2.0 x16
Molex Power Connectors 2 x 6-pin  1 x 6-pin, 1 x 8-pin 2 x 6-pin 1 x 6-pin, 1 x 8-pin 1 x 6-pin, 1 x 8-pin
Multi GPU Technology SLI  CrossFireX CrossFireX SLI CrossFireX
DVI Output Support 2 x Dual-Link  2 x Dual-Link 2 x Dual-Link 2 x Dual-Link 2 x Dual-Link
HDMI 1 1 1  1 (mini-HDMI) 1
DisplayPort 1 (version 1.2) 2 (version 1.2 HBR2) 2 (version 1.2 HBR2) None 2 (version 1.2)
HDCP Output Support Yes  Yes Yes Yes Yes
Street Price Launch Price: US$499  Launch Price: US$549 ~US$449 ~US$509 ~US$349


The GeForce GTX 680 has been designed for quieter gaming and when it was under load during our benchmark testing, the GPU fan was visibly quiet even when fan speed was set at 60%. Do refer to our test setup section of the article for more details on the usage of EVGA Precision, a third party utility for tweaking the GTX 680 graphics card.

The card features two Dual-Link DVI ports, a HDMI port and a DisplayPort for support of up to four displays - all natively on the GPU.

The 6-pin power connectors are staggered in position so that the release pins of the power plugs from the PSU are not aligned. This allows for easier removal of the power plugs from the graphics card. However, the design also hints of an even faster GPU variant that could take up yet another power socket with the spare space; time will tell if we're right.

The Good
Excellent all-round performance
Extremely quiet for a flagship card
GPU Boost in Kepler architecture
The Bad
Temperature and power consumption wasn't low but was justified with GPU boost