Nvidia GeForce GTX 980 4GB “Maxwell” (28nm) Review
What’s New?
VGXI (Voxel Global Illumination)
Nvidia has an extended history of launching new graphics technologies alongside new product releases and the GTX 980 isn’t an exception. Three new Nvidia technologies launch alongside the GTX 980, the first of which is VXGI (Voxel Global Illumination). VXGI is a truly dynamic global illumination system that Nvidia has been working on since 2011. It uses a voxel grid and cone tracing to calculate an approximation of global illumination and ambient occlusion in real time.
“VXGI includes support for a number of new graphics features that dramatically speed up the voxelization process where we convert triangles into voxels. For instance, conservative raster is used to quickly test which pixels touch the triangle so we can accurately convert the 3D scene geometry into voxels. VXGI also features a multi-projection engine that takes the geometry and re-projects it onto multiple surfaces simultaneously.”
DSR (Dynamic Super Resolution)
DSR, or Dynamic Super Resolution, is Nvidia’s fancy way of talking about downsampling. The GPU renders the game at a higher resolution, up to 4K in fact, then down-samples it to fit on a 1080p screen. The overall effect is even though the same resolution is presented, the visual quality of the game is actually increased. I’m not sure how it works, but it does work! Nvidia explains DSR as follows:
Dynamic Super Resolution is similar to the downsampling method many eye candy purists are using today, although many games haven’t been designed with downsampling in mind, leading to artifacts in the final rendered image.
To address this issue, Dynamic Super Resolution uses a 13-tap Gaussian filter during the conversion to display resolution. This high-quality filter reduces or eliminates the aliasing artifacts experienced with downsampling, which relies on a simpler box filter.
Another issue with conventional downsampling is ease of use. Because it isn’t a native solution, downsampling requires users to trick the GPU into thinking it’s connected to a higher resolution display than it actually is. In order to do this, users have to create custom screen resolutions and often have to adjust various low-level display parameters in order for downsampling to work.
In comparison, Dynamic Super Resolution can be found inside GeForce Experience, where we provide easy to use controls for enabling Dynamic Super Resolution with optimized game settings – simply click the “Optimize” button and you’re done!
MFAA (Multi-Frame Sampled AA)
Nvidia has always pushed the boundaries of anti-aliasing technologies; TXAA and MSAA are recent examples and Nvidia’s latest effort is being called MFAA. MFAA (Multi-frame Sampled AA) provides additional resource reductions compared to MSAA. MFAA is targeted at 4K resolutions where traditional anti-aliasing technologies cripple frame rates.
Multisample anti-aliasing (MSAA) is the most common form of AA in use by gamers today. But MSAA can be expensive from a performance perspective, especially once you scale to higher resolutions like 4K. Many graphically demanding games can’t deliver playable frame rates at 4K with 4xMSAA applied.
To overcome this issue, NVIDIA engineers have developed MFAA. Utilizing a new hardware feature found in the GeForce GTX 980/970 (programmable AA sample positions), MFAA alternates between multiple AA sample patterns to produce the best image quality while still offering a performance advantage compared to traditional MSAA. In our 4xMSAA example, MFAA alternates between two different 2xAA sample patterns that are blended together with a temporal synthesis filter to produce an image that looks similar to 4xMSAA to your eye.
The GTX 970 is a far better buy. It costs on average 30% less and sacrifices only ~18% in performance.
The new Maxwell architecture really seems to be uber efficient doing a lot more with a lot less. Impressive stuff indeed.
Sadly our GTX 970 only arrived four hours after the launch, the review will be coming next week though 🙂
I’ll wait to hear what you have to say but I’m confident you won’t be disappointed. 😉
“Amd was already quite some way behind kepler in power efficiency”, that’s not a true statement I’m afraid. The 780 ti and 290x performs almost equally while the 290x draws slightly more power. If you wanna make a statement how about explaining why it takes nvidia a 563mm2 die to compete with amd’s 439mm2 die and only just match it in real world performance. Also tell us why the fully unlocked kepler can only muster 1.7 tflops of double precision while full hawaii is capable of 2.8 tflops, Nvidia’s software allows them to compete with amd’s superiot architecture It seems.
I couldn’t agree more, but it seems that the new AMD cards are not totally new. I mean, starting from the HD7900 series, AMD hasn’t change the entire line. This is not good, so we should wait for another year. In any case, AMD R7 370 is a really good card (if you are in Europe, you may look for it already – http://hardware.nl/sapphire/11240-06-20g.html but I have read here http://www.pcworld.com/article/2976945/components-graphics/amd-quietly-launches-new-radeon-r9-370x-graphics-card-but-not-in-the-usyet.html that there are the USA customers cannot get it yet. Strange.