NVIDIA made the surprise announcement of GeForce GTX 690 launch last weekend. A little later, we found ourselves opening a classy wooden crate with a crowbar NVIDIA sent a little earlier, in it was the GeForce GTX 690. The GTX 690 is NVIDIA's quest for absolute performance leadership in the graphics processor market. Despite being months behind AMD in launching its 28 nm Kepler GPU lineup, NVIDIA caught up with it, in the performance segment, with the GeForce GTX 680, and now it seeks to settle the disputed absolute performance lead between GeForce GTX 590 and Radeon HD 6990, with the new GeForce GTX 690.
In more ways than one, this launch is an assertion of NVIDIA's technological leadership, because AMD still hasn't launched its enthusiast-segment dual-GPU graphics card, yet, and one can't expect it to be out before June. The GeForce GTX 690 is a dual-GPU graphics card with two 28 nm GK104 GPUs, the same chips found in the GeForce GTX 680. The two chips each are clocked slightly lower than GeForce GTX 680, and rely on a PCI-Express 3.0 bridge chip for bus interface, however, NVIDIA claims that GTX 690 should provide performance comparable to GeForce GTX 680 2-way SLI. The card also costs the same as two GTX 680s, at US $999. Which makes it the most expensive reference design graphics card ever released.
The GTX 690 has two GK104 Kepler GPUs arranged in an internal SLI configuration. Each GK104 chip has all its components enabled, including 1,536 CUDA cores, 128 TMUs, 32 ROPs, and four independent geometry processing units. Each GPU further has a 256-bit wide GDDR5 memory interface, with which it talks to 2 GB of memory. The total memory on the card, hence is 4 GB. In this GeForce GTX 690 review, we verify NVIDIA's performance claims (of it matching 2x GeForce GTX 680), because that is key to ascertaining whether the card is worth $999, or if NVIDIA is already having the spoils of war.
|Shader Units||512||1792||2048||1536||2x 1536||2x 512||2x 1536|
|ROPs||48||32||32||32||2x 32||2x 48||2x 32|
|Graphics Processor||GF110||Tahiti||Tahiti||GK104||2x Cayman||2x GF110||2x GK104|
|Transistors||3000M||4310M||4310M||3500M||2x 2640M||2x 3000M||2x 3500M|
|Memory Size||1536 MB||3072 MB||3072 MB||2048 MB||2x 2048 MB||2x 1536 MB||2x 2048 MB|
|Memory Bus Width||384 bit||384 bit||384 bit||256 bit||2x 256 bit||2x 384 bit||2x 256 bit|
|Core Clock||772 MHz||800 MHz||925 MHz||1006 MHz+||830 MHz||607 MHz||915 MHz+|
|Memory Clock||1002 MHz||1250 MHz||1375 MHz||1502 MHz||1250 MHz||855 MHz||1502 MHz|
We received the card from NVIDIA in a wooden crate that had to be opened using a crowbar.
Retail boards will probably use a different form of packaging and certainly come with the usual accessories.
NVIDIA engineered a beatiful cooling solution around their dual GPU card. It uses lots of metal and plexiglas. For me this is one of the most sexy and elegant coolers out there.
NVIDIA's card requires two slots in your system.
Display connectivity options include three dual-link DVI ports, and one mini DisplayPort. You may use all the outputs at the same time, thanks to NVIDIA Kepler's new display output controller.
An HDMI sound device is included in the GPU, too. It is HDMI 1.4a compatible which includes HD audio and support for Blu-ray 3D movies. The DisplayPort outputs are version 1.2 which enables the use of hubs and Multi-Stream transport.
You may combine up to two GTX 690 cards from any vendor in a quad-GPU SLI configuration for higher framerates or better image quality settings.
Pictured above are photos of the front and back, showing the disassembled board. High-res versions are also available (front, back). If you choose to use these images for voltmods etc., please include a link back to this site or let us post your article.
Our Patreon Silver Supporters can read articles in single-page format.