Published in Graphics

Nvidia 20nm Maxwell comes in late 2014

by on09 April 2014

The real DirectX 12 part

As we have said before Microsoft’s new DirectX 12 API will be supported by Fermi, Kepler and Maxwell products, but it turns out that only the new Maxwell chips will support the DirectX 12 hardware feature set. Nvidia is working hard on its 20nm Maxwell chips. These are different from the first generation Maxwell that you can get today in Geforce GTX 750 cards. The first generation Maxwell was codenamed GM107 and it is a 28nm part.

It was quite logical that Nvidia would push for a 20nm transition in this year, as it turns out that the wafer economy and performance per watt will finally make sense. The 20nm process has been around for quite some time, but there is a clear difference between test production and volume production.

We expect to see 20nm Maxwell with DirectX 12 hardware support in the second half of 2014, the latter part of the year to be more specific. However, nothing is set in stone yet. We are not aware if AMD can launch its 20nm chips counterpart at the same time, as some info we gathered implies that AMD’s 20nm high end chips with full DirectX 12 hardware support might be coming in early 2015. This wouldn’t be surprising, either, as Hawaii is relatively fresh and it can easily hold its ground until 20nm Maxwell parts show up.

Since DirectX 12 hardware supported games are expected to launch in time for the 2015 holiday season 2015, both AMD and Nvidia have quite a bit of time to get their DirectX 12 hardware ready, but once again once you move from 28nm to 20nm, you can expect significantly better thermals, higher clocks as well as more transistors on the same die size. DirectX 12 is just the icing on the cake.

If Maxwell is ever to reach between 8 and 16 Gigaflops performance as outlined in Nvidia’s roadmap, the upcoming parts need to pack more transistors. Currently the fastest single chip Geforce GTX Titan Black has 5,121 Gigaflops and we expect that an upcoming high-end GK chip will start with at least 8 Gigaflops single precision, as this is what Nvidia GPU roadmap has implied.

It will take a dual card and probably a lot of time until Nvidia’s first 20nm GM architecture reaches 12 Gigaflops.

Rate this item
(0 votes)

Read more about: