NVIDIA Tegra Architecture

I like that the modem can be updated in the field to new standards.
I thought that was the same with other modern chips too (Qualcomm has a couple of Hexagon cores in their basebands).
In any case though I wish you luck getting any updated modem firmware supporting some standard which wasn't promised already when shipping. Even if it would be possible, chances of that happening are imho zero.

Oh, that's a separate modem and not use of a Tegra 4i.
That was not entirely clear to me.
Well that would otherwise be quite a downgrade, I don't think customers would like that (get either slow, more expensive tablet with LTE or fast tablet without...). AFAIK the modem is exactly the same as in tegra 4i it's just a separate chip.
 
Yeah realistically, they would want you to buy a new device and would hype up the new device as being compatible with the faster new network.
 
Firmware for a software defined radio has to be REALLY locked down, I wonder if it's practical or not to update it over the internet. That's a rhetorical question I'm afraid.
It's a similar problem to updating a car's critical firmwares, you don't want Joe user or malware authors/attackers messing with it nor do you (or the FCC and similar) want the public be able to mess with the cell networks, emergency and military communications, or just use it to poison the spectrum.
 
Even if the baseband firmware could be updated to add new cellular standards I imagine you're still limited by supporting hardware like antennae and power amps not being designed for it.
 
I like that the modem can be updated in the field to new standards.

The benefit to NVIDIA with their software-defined modem is that they do not need to re-architect their modem to support more advanced features. For instance, moving from LTE Cat 3 to LTE Cat 4 with carrier aggregation simply requires a software update. This reduces cost and enhances time-to-market (all else equal). The downside is that NVIDIA cannot trumpet up any physical changes to their modem because the changes are being made in software.

One of NVIDIA's biggest challenges is to convince partners to use the i500 modem. The Tegra 4 generation was and is important because it proved that i500 has the reliability, performance, and feature set to pass the most stringent testing on the AT&T and Vodafone voice and data networks.
 
Last edited by a moderator:
Even if the baseband firmware could be updated to add new cellular standards I imagine you're still limited by supporting hardware like antennae and power amps not being designed for it.

The selling point of updatable baseband firmware isn't for the consumer. It's to allow quick adoption of new network standards, something that NVIDIA and the OEMs they sell to would/could benefit from.
 
First Tegra 4i phone priced at €200

http://www.fudzilla.com/home/item/34289-first-tegra-4i-phone-priced-at-€200

It has a 4.7-inch 720p display, 8-megapixel camera backed by a 5-megapixel front facing camera. It has 1GB of RAM, but only 4GB of on board storage. Luckily there's a microSD slot for expansion. All in all it looks like a good deal for €200. That sort of money will buy you a Galaxy S3 mini, a Huawei G610 or a Moto G. Although we've huge fans of the latter, none of them come close to the Wiko Wax in terms of specs and sheer power. None of them has LTE, either.
 
NVIDIA Posts Working "GK20A" Support For Nouveau

Alexandre Courbot of NVIDIA published these patches for providing initial hardware support for the GK20A graphics core inside of Nouveau
The GK20A is the Kepler-based graphics core found within the Tegra K1, NVIDIA's very powerful next-generation ARM SoC that will eventually feature 64-bit ARM cores and a graphics core derived from their mainline Kepler architecture. While there's been separate work in the past on Tegra "GeForce" graphics, with Tegra K1 and beyond they're on the same architecture as the desktop GeForce graphics and can be relying upon the same driver code-base.
 
hrHJ0ht.jpg


aSYWXQY.jpg


http://nvidianews.nvidia.com/Releas...e-Supercomputer-for-Embedded-Systems-ad8.aspx
 
Last edited by a moderator:
So NVIDIA showed a slide showing Tegra K1 which mentioned 326 GFLOPS (if I saw it correctly) instead of the 365 (GPU) GFLOPS on the original announcement presentation. Any reason for this difference?

EDIT: Jetson has 326 GFLOPS, so that's probably where the number comes from.
 
So NVIDIA showed a slide showing Tegra K1 which mentioned 326 GFLOPS (if I saw it correctly) instead of the 365 (GPU) GFLOPS on the original announcement presentation. Any reason for this difference?

EDIT: Jetson has 326 GFLOPS, so that's probably where the number comes from.

If the actively cooled Jetson has 326 GFLOPs, I wonder what a smartphone / tablet will actually ship with, less than 200 GFLOPS?
 
Back
Top