At this year’s GTC – GPU Technology Conference held from March 24th till March 27th, Nvidia had made some very important announcements and demonstrations. One of these was the announcement of the second generation Maxwell, based on the 20nm lithography process. The 20nm Maxwell is expected to have double the performance per watt of the current generation. It is also to be integrated with Project Denver, thereby fusing general purpose ARM cores alongside the GPU. Cards based on the architecture were expected to be released somewhere around 2014, but we recently got word of their delay. More news is in however, and this time it relates to next generation display capabilities i.e the 4K UHD.

HEVC-vs-H2641

Confirm Support in Next Gen Maxwell for H .265. Google’s VP9 Codec Support Not Confirmed.

As we know, H .265 and VP9 both are competing Codecs for the future. Both of these claim twice the efficiency of the current H.264. Both reduce file size significantly, making stream and transferring of content much easier and smooth. H.265 increases pixel size, which in turn reduces bitrate and therefore reduces overall size of the file. H.265 takes 64 x 64 ‘superblocks’ and performs 35 infra-prediction directions to rebuild them. VP9 works in a relatively similar way as well. The only major difference here, apart from VP9 being open source and royalty free is that the latter does not require the blocks to be square. They can have a variety of dimensions for example 4 x 8 or 64 x 32. However, it only has the capability to perform 10 prediction models for rebuilding. Nevertheless, both of these standards require significantly greater computational power when compared to their predecessors.

According to our friends over at Fudzilla, who have learned this from industry sources, the 20 nm Maxwell is going to support H.265. This is no surprise though, as this definitely sounds to be the codec for the future. No word was received on potential support for VP9, however it would not be surprising if it were included as well. Both the codecs can support up to 8K resolution. Maxwell’s next generation is expected to have significantly greater computational power when compared to the present one and it is on the basis of this power that it will be able to support 4K, which has four times the pixels of current 1080p resolution. AMD and Intel won’t be the ones to lag behind in this support as well. However the scenario turns out, one thing is for sure. We will most certainly get much, much higher quality content which will stream much, much quickly to our devices in the future. We can’t wait, can you?

NVIDIA GeForce Maxwell Details Analysis:

Several details of the GeForce Maxwell chips have leaked in the last  few weeks ago listing several high-end GeForce chips such as the GM204 and GK210. We came to the conclusion that these chips will be configured for different tiers of cards that are mentioned below:

  • NVIDIA GM200 (Maxwell Architecture, High-Performance for Telsa/Quadro Arrives later for Cosnumers, Successor of GK110)
  • NVIDIA GM204 (Maxwell Architecture, High-End Consumer, Successor of GK104, First GeForce 800 Series Products likely to feature)
  • NVIDIA GM206 (Maxwell Architecture, Performance Minded, Successor of GK206, Mid-Range GeForce 800 Series products to feature)
  • NVIDIA GM107/207 (Maxwell Architecture, Entry Level, Successor of GK107, Entry Level GeForce 800/700 Series To feature, Already introduced on GTX 750 Ti / GTX 750)

One thing that still hasn’t been confirmed it whether or not these chips will feature a 20nm process but its most likely that NVIDIA will move for 20nm now considering that their GM107 series based on the 28nm process has already been out for a while and we will be approaching 2H of 2014 soon. The GeForce 800 series will include a flagship part that’s not based on the GK110′s successor (GM200/GM210) but rather the GM204 which replaces the GK104. NVIDIA launched their first Kepler high-performance card with the GK104 core so its probably that NVIDIA may use GM204 to power their GeForce GTX 880 graphics card. The Maxwell GPU architecture will be fully compatible with DirectX 12 featuring low-CPU overhead in addition to new features that are only going to be introduced on next generation chips such as Maxwell itself.

The memory bus is quite tricky to confirm right now but 512-bit seems highly plausible coupled with 8 GB of GDDR5 memory while the high-end GK110 replacement would be either 512-bit or greater.

Advertisements

GeForce GTX 480 GeForce GTX 580 GeForce GTX 680 GeForce GTX 780 GeForce GTX 880 GeForce GTX 980
Codename GF100 GF110 GK104 GK110 GM204 GM200
Process 40nm 40nm 28nm 28nm 28nm/20nm 20nm
GPU Core Fermi Fermi Kepler Kepler Maxwell Maxwell
Memory Bus 512-bit 512-bit 256-bit 512-bit 256/512-bit 512-bit
Launch March 26th 2010 November 09 2010 March 22nd 2012 May 23rd 2013 2H 2014 2015
Price $499 US $499 US $499 US $499 US $499 US? $499 US?

 

We cater to your constant need to remain up to date on today’s technology. Like us, tweet to us or +1 us, to keep up with our round the clock updates, reviews, guides and more.