Nvidia hasn't said exactly when and how it will unleash its next-gen Maxwell graphics architecture. However, a leaked roadmap unearthed by Computer Base may hold some clues.
The roadmap is from Clevo, a maker of notebook PCs, and it can still be downloaded from PCW.fr, a French distributor of Clevo notebooks. While the document doesn't mention Maxwell specifically, it does suggest that a lineup of GeForce GTX 800M-series mobile GPUs is due out next February:
Beside quoting memory capacities (which will apparently range up to an impressive 8GB for the GeForce GTX 880M), the roadmap says the GeForce GTX 860M, GTX 870M, and GTX 880M will be available as discrete MXM cards. Another slide mentions three "on board" offerings: the GTX 840M, GTX 850M, and GTX 860M. Computer Base suspects "on board" simply means soldered-on. If that's true, then there will be both MXM and BGA versions of the GTX 860M.
Of course, none of this tells us what graphics architecture will make the GTX 800M series tick. These could simply be re-badged Kepler chips. Going by the memory capacities alone, though, I wouldn't be surprised if there were a little next-gen goodness inside the GTX 870M and GTX 880M. 6-8GB is a lot of memory for a consumer-focused mobile GPU. Nvidia's current mobile flagship, the GeForce GTX 780M, comes with 4GB of GDDR5, and I'm not sure Nvidia would have much to gain from giving the same silicon even more RAM to play with.
(Thanks to TR reader SH SOTN for the heads up on this one.)
|The TR Podcast 175: the Zen of chipmaking and ARM's Cortex-A72 revealed||4|
|Elon Musk lays out vision for a battery-powered future||103|
|Inside ARM's Cortex-A72 microarchitecture||34|
|Asus' 144Hz MG279Q monitor may top out at 90Hz with FreeSync||56|
|Deal of the week: A Bay Trail netbook for $161, free case fans, and more||18|
|DirectX 12 Multiadapter shares work between discrete, integrated GPUs||95|
|Gigabyte's 9-series motherboards are Broadwell-ready||44|
|The TR Podcast will be live on Twitch shortly!||3|
|AMD delays FreeSync support for multi-GPU systems||40|