It's not an 'Ultra'book without Kepler, says NVIDIA

by Alistair Lowe on 24 April 2012, 11:14

Tags: NVIDIA (NASDAQ:NVDA)

Quick Link: HEXUS.net/qabfoj

Add to My Vault: x

Following the release of Intel's Ivy-Bridge processors, NVIDIA has been quick to remind us all that, even with Intel's new integrated HD 4000 graphics core, which we benchmarked earlier, an Ultrabook isn't 'Ultra' without Kepler and has thus announced its mobile line-up.

There's a little confusion and conjecture as to the origins of the firm's mobile GPUs, with the specifications for some indicating that they are still based on the now, two-year-old Fermi architecture, produced at 40nm. It had been suggested quite some time ago that 28nm production at TSMC was at its limit and, that this could be one of the primary motivations behind NVIDIA's mixed release.

The firm has clearly chosen to focus its Kepler-based cards at the lower-end, where power-efficiency is paramount, with the 640M and 640M LE specifically targeting Ultrabooks. Quite rightfully holding off on the release of a 680M, it's clear NVIDIA intends to introduce a Kepler-based pack leader in the not too distant future.



HEXUS Forums :: 5 Comments

Login with Forum Account

Don't have an account? Register today!
the 28nm kepler cards look interesting but “up to 384” shaders is not precise enough.

i want to know what i am buying, and i don't like non-functional units!

is it only the 660M that has 384 shaders?
does even the 660M have all 384 shaders?
Interesting just how many of those are still 40nm. I know that 28nm isn't mature yet, but I'd still have expected more of a push there on the mobile front.
Do product numbers mean nothing anymore?
I'm sure they do say that. But really.. if you want long battery life, you'll get a machine without a discrete GPU. You're not going to buy one of these things for high-end gaming or extreme parallel computation anyway.
Yet their top end parts aren't Kepler at all?