Nvidia Demos GeForce "Kepler" in Action as Photos of GeForce GTX 670 Ti Hit the Web.
The Mystery Behind Nvidia's Kepler Begins to Clear Away: 2-3 Times Faster than Fermi[03/08/2012 07:49 PM]
by Anton Shilov
As Nvidia Corp. demonstrated its next-generation GeForce code-named Kepler in action at Game Developers Conference (GDC) and got applauds for impressive performance, a Chinese web-site published an image of what it claims to be GeForce GTX 670 Ti graphics card.
Epic Games, a major game designer, astonished the GDC12 attendees with its Samaritan demo, which depicts next-generation video-game graphics effects using Nvidia GeForce GTX "Kepler" hardware. Created in a dark, futuristic setting, the demo utilized a host of advanced rendering techniques that smoothly tessellated and morphed facial features, created realistic street scenes using point light reflections, and replicated the work of the best movie directors through the use of fine-tuned out of focus bokeh filters.
Back at GTC11, a similar demo was unoptimized and was run on three Nvidia GeForce GTX 580 graphics boards in multi-GPU mode. This time, however, at GDC12, the Samaritan (which is based on the tweaked Unreal Engine 3 demo) was rendered using one Nvidia GeForce GTX "Kepler" graphics cards. Potentially, this proves that the next-gen Kepler architecture is two or more times faster than the existing Fermi architecture in its model 580 form. The demo video is available below.
In the meantime, RedQuasar web-site has published an image of a graphics card that carries the GeForce GTX 670 Ti moniker and which is supposed to be one of the first Kepler GPU-based offerings.
The board has two MIO connectors, which are found only on high-end products, to provide three-way or four-way multi-GPU SLI configurations. The new adapter has DisplayPort, two dual-link DVI-I and HDMI display connectors. In addition, the card seems to have tricky PCIe power connectors previously found on an alleged Kepler-class graphics board.
Nvidia has very high expectations for its Kepler generation of graphics processing units (GPUs). The company claims that it had signed contracts to supply mobile versions of GeForce “Kepler†chips with every single PC OEM in the world. In fact, Nvidia says Kepler is the best graphics processor ever designed by the company. Unfortunately for Nvidia, yields of chips made using 28nm process technology at TSMC, including Kepler, are lower than the company originally anticipated and therefore their costs are high and Nvidia's chief executive recently complained that the company cannot get enough wafers.
Nvidia GeForce "Kepler" GK104 graphics card looks exactly like a high-end adapter: it has very tricky 5-phase power supply system with very unusual set of PCIe power connectors. According to some media sources, the external power supply system may include two 6-pin and an 8-pin PCIe power plugs, which means that the card will have rather high power consumption. -- Efficiency goes out the window again
The GK104 chip does not seem to be too huge and therefore too expensive. Some believe that Nvidia is also preparing code-named GK100 chip, which will have even higher graphics and compute performance. Exact specifications of the GK104 are unknown. -- One month from launch and still no hard facts
Nvidia's official GTX 680 performance slide has leaked online and it looks like Nvidia is quite confident, perhaps even cocky. According to the slide, Nvidia's GTX 680 is over 40 percent faster than HD 7970 in Battlefield 3.
Of course, we are talking about an Nvidia slide here and this is most likely the best case scenario, to put it mildly. There are no details regarding drivers that were used or even resolution on which these couple of games were tested. The most impressive lead scores of over 40 percent and over 20 percent are in Battlefield 3 game when 4xAA is used and in Dirt 3 with and without AA.
According to what we heard earlier, the GTX 680 is indeed faster but only in some games. Of course, we just need to wait for some official benchmarks that will put these bold Nvidia claims to test. The slide came from NGF Community forum.
We already wrote that Nvidia planned to hold a GTX 680 related event yesterday, and it is no wonder that we now have a bit more info regarding the card, while Chiphell has provided a much more clear picture of the actual card.
It looks like our sources were spot on yesterday when we heard that the new GTX 680 will feature two 6-pin PCI-Express connectors and have a TDP of around 200W (190W to be precise). When combined with 1536 shader units, 2GB of GDDR5 memory clocked at 6GHz and, according to what we heard, a very high GPU clock, it comes as no surprise that Nvidia is claiming higher performance per Watt for its upcoming GTX 680.
This is a big win for Nvidia as we aren't used to hearing performance per Watt from Nvidia and now it appears that the tide has turned. The performance of the GTX 680 is somewhere around Radeon HD 7970 as it wins in some games and benchmarks and loses in others. Of course, these are all results that came from Nvidia so we'll hold our judgement until we see some reviews.
As noted before, the card is based on Nvidia 28nm GK104 Kepler chip that is around 300mm² in size.
The bad part of the story is that our sources were again spot on with the US $549 price tag. Nvidia has a card with similar performance and better TDP so the US $549 price tag isn't really a surprise. We guess that Nvidia has additional maneuvering room when it comes to price and we just need to wait and see if AMD is going to bite down and head into a new round of price wars.
NVIDIA's New AA Algo is TXAA, Adaptive V-Sync and New 3DVision Surround Detailed[font=Verdana, Arial, Helvetica, sans-serif]
With Kepler architecture, NVIDIA has three new star technologies that will help it with this round: TXAA, a new anti-aliasing algorithm that offers image quality comparable to 16X MSAA, with the performance-penalty of 2X MSAA (if not less); Adaptive V-Sync which is sure to win gamers by the millions; and a redesigned display logic that supports up to four displays from a single GPU.
TXAA, which we talked about a little earlier, turns out to be a super-efficient temporal anti-aliasing algorithm. It has two levels: TXAA(1), and TXAA2. TXAA1 provides the image quality comparable to 16X MSAA, with the performance-penalty of 2X MSAA; while TXAA2 offers image quality higher than 16X MSAA (unlike anything you've seen), with the performance-penalty of 4X MSAA. Since few games natively support it, you will be able to enable it through the NVIDIA Control Panel, in the application profiles, provided you have a Kepler architecture GPU.
More pictures follow.
Adaptive V-Sync is a smart frame-rate limiter that fluidly adjusts frame-rate when heavy 3D scenes drop frame-rate below monitor refresh-rate and above it. When frame-rate is lower than monitor refresh-rate, lag occurs, and overclocking attempts to overcome it. When frame-rate is higher than monitor refresh-rate, page-tearing occurs, and the normal functionality of V-Sync takes over. Adaptive V-Sync makes these transitions "organic".
Then there's the new 3D Vision Surround, bolstered by a redesigned display logic, which addresses the two-display limitation of NVIDIA GPUs. You can now connect as many as four monitors to a GeForce Kepler GPU, enabling 3-monitor HD 3D Vision Surround setups. You no longer need more than one GeForce GPU to connect more than two monitors. The new 3D Vision Surround is said to work in conjunction with Adaptive V-Sync to ensure the center display has higher frame-rate (since it's at the focus of your central vision), at the expense of the frame-rates of the two side displays (since they're mostly at your peripheral vision). This ensures there's a balanced, high-performance experience with multi-monitor gaming setups.
[/font]