Traditionally, graphics cards launches follow a set order. Usually the flagship single GPU solution is released and then followed by the more affordable graphics cards in the stack. However, Nvidia is turning this on its head by releasing the GTX 690 just a month after the GTX 680 launched.
Making today’s announcement even more exciting is the GTX 690 is the most powerful and refined graphics card Nvidia has built to date. This is clear just for the exterior of the graphics card, as it doesn’t use plastic casing or generic materials. Instead it is utilizes a thermal solution constructed of cast aluminum and magnesium alloy with polycarbonate windows and nickel plating, making it the most expensive and solidly constructed graphics card Nvidia has ever built. Beyond the exterior though the GTX 690 is also built on a 10 layer 2oz PCB, has ample 10 phase power supply and two GK-104 graphics processors.
By now most of us are familiar with the GTX 680, and the new features such as Nvidia GPU Boost technology, which is a major factor in the performance of the GTX 690. According to Nvidia the GTX 690 is able to perform nearly 2x faster than the single GPU based GTX 680. This is due to the power efficiency of the Kepler architecture, which allows Nvidia to keep the Boost speed of the dual cores above 1GHz, while staying in an acceptable power envelope. The power efficiency, along with the robust thermal solution also allows the GTX 690 to run at a low acoustic level of only 47dBA, making it as quiet as many case fans.
The GTX 690 also utilizes binned graphics processors that were handpicked for their low leakage and power efficiency. Essentially, this means Nvidia has selected the best of the best graphics processors, which allowed them to keep the base and boost clock at 915MHz and 1019MHz respectively. But like the GTX 680 the Boost clock is just an estimate and it should run at faster than 1015MHz during gaming. In addition the GTX 690 has dual Kepler cores that internally feature 3072 CUDA cores, 64 ROPs, 128 Texture units and a 4GB GDDR5 frame buffer running on a 512-bit interface. So, Nvidia didn’t scale anything back, as the GTX 690 packs in two full GK-104 graphics processors.
So, what is this beast going to cost? Well... at $999 it is double the cost of a single GTX 680. The tradeoff though is it is a single PCB solution that is quieter than dual GTX 680s and comes with massive bragging rights.
Along with the GTX 690, Nvidia also introduced a new program called the GeForce Experience. Essentially, GeForce Experience has been a long term program where Nvidia ran programs and games through a super computer for it to find the best settings for the GPU and CPU configuration being used. When being used the program allows a host system to connect to the super computer allowing gamers that traditionally don’t change the in game settings to enjoy the maximum eye candy available by automatically allowing the program to adjust the settings. There is some user interaction still, but it is only limited to hitting an optimize button, and GeForce Experience takes care of the rest.
Be sure to check back May 3rd for our full launch review.