NVIDIA shows off new roadmaps for graphics and mobile chips

When he wasn't praying to "Outlook God" for a Windows RT port of Outlook, NVIDIA CEO Jen-Hsun Huang was presenting some new information on Tuesday about the company's future plans for graphics and mobile processors as part of NVIDIA's annual GPU Technology Conference.

NVIDIA's blog offers up the details of the presentation. In 2014, Huang stated that Maxwell will be launched as the successor to NVIDIA's current Kepler design for its GeForce graphics chips. Huang said it will use "unified virtual memory, which makes it possible for GPU operations to see the CPU memory and vice versa."

After that comes Volta, which NVIDIA will introduce in 2015. Volta will use something called "stacked DRAM" which Huang said " ... will have DRAM on same silica sub-strate, which will carry a whole bunch of DRAMs stacked atop each other. We’ll cut a hole through the silicon and connect each layer. We’re going to achieve one terabyte per second of bandwidth."

As far as its Tegra mobile chip lineup, NVIDIA promises to bring the successor to Tegra 4, code name 'Logan', in 2014. Huang said, "It has a Kepler GPU, with full CUDA 5, and OpenGL 4.3." In 2015 we get the next version of Tegra, code named 'Parker', which will be NVIDIA's first 64-bit ARM processor and will use the Maxwell graphics design.

In 2012, NVIDIA announced the GeForce Experience, a cloud-based system that allows PC gamers to have the best graphics settings on their games based on data from NVIDIA's servers. Today, NVIDIA announced its first cloud-based server for enterprise customers, the GRID Visual Computing Appliance. It's NVIDIA's first server product and each GRID VCA has two high end Xeon processors inside, along with 16 Kepler-based GPUs. Each box can support up to 16 users.

Source: NVIDIA | Images via NVIDIA

Report a problem with article
Previous Story

Interview with the CEO of Windows 8 app analytics company MarkedUp

Next Story

Microsoft team members targeted in Xbox Live account attacks

11 Comments

Commenting is disabled on this article.

"access to the CPU Memory, and vice verse": Does this mean that the CPU will be able to directly access the GPU's memory (meaning your processor will have ~1/1.5/2/3 GB of....L4....cache?) I know it's not L4, just using that as next in line after L3. Or am I misunderstanding this?

I must have been the only one who read this and immediately felt disappointment with the code names for Tegra.

They should have stayed in the DC universe.

cpu killer said,
Seems like I'm solid with the 7950 for a while longer. Hopefully the next generation brings some awesome improvements.

both current gen AMD and NVIDIA will be good for a while, both companies are churning out new drivers that are taking performance of their cards to new levels evey few months.
It's good that they took an odd year for software only optimizations instead of trying to throw more hardware towards their not-so-optimal drivers.

Fus10n said,
Yes! I have been waiting to upgrade to 780s!

EDIT.. Didn't realize it was going to take so long.. ugh

I'd imagine a kepler refresh will come out in 2013 and be the 780. Maxwell will most likely be 880.

Yea, I"m kinda impatient for 780s, even if they aren't a new gen in themselves, since I'm on a 580 now and can't justify upgrading now to 680...