facebook rss twitter

Tegra 2 flexes its muscles

by Pete Mason on 31 August 2010, 10:18

Tags: NVIDIA Tegra, Toshiba (TYO:6502), NVIDIA (NASDAQ:NVDA)

Quick Link: HEXUS.net/qazsu

Add to My Vault: x

NVIDIA has been touting the performance of its long-awaited Tegra 2 SoC for some time, promising a 'revolution' in the portable market.  However, a benchmark released over the weekend shows that the upcoming chip might actually blow the competition away.

The folks over at Carrypad managed to get their hands on a Tegra 2-powered Toshiba AC100 netbook running the new silicon, so naturally decided to put it through its paces.  Frankly, the results speak for themselves, with the upcoming platform beating other recent mobile-chipsets by a significant margin in the Quadrant benchmark.

The source also reported impressive speeds across a number of other tests, exceptionally smooth video playback and a generally fluid experience when navigating the OS.

Furthermore, the AC100 was running Android 2.1, as opposed to Android 2.2 (‘Froyo'), which was released earlier this year.  The newer version of the OS is known to bring a significant performance-increase, and could see the Tegra 2-platform walk even further away from its rivals. 

The follow up to NVIDIA's powerful first-generation SoC is a dual-core media-processor that will reportedly be powering a number of upcoming devices.  Not only will this include phones, but a good number of tablets - including one from Toshiba - are rumoured to be making use of the chip, as well netbooks like the AC100.

Benchmarking on phones is much less mature - and reliable - than on PCs and some of the test suites are known to give anomalous results, especially when running tweaked kernels.  However, with the netbook reported to be running an unmodified version of Android, the Tegra 2 could easily live up to NVIDIA's claims. 

HEXUS Forums :: 4 Comments

Login with Forum Account

Don't have an account? Register today!
Posted by Biscuit - Tue 31 Aug 2010 14:43
But does it come at the expense of battery life… that is the real question. Performance figures mean very little in mobile devices when they are not compared to power usage.
Posted by kingpotnoodle - Tue 31 Aug 2010 15:20
Biscuit
But does it come at the expense of battery life… that is the real question. Performance figures mean very little in mobile devices when they are not compared to power usage.

Thats what I thought…. tablet/netbook vs phone is hardly a fair comparison, tablet/netbooks have far bigger batteries and can hold and cool far more CPU.
Posted by BullDogg - Tue 31 Aug 2010 15:29
kingpotnoodle
Thats what I thought…. tablet/netbook vs phone is hardly a fair comparison, tablet/netbooks have far bigger batteries and can hold and cool far more CPU.

But the Tegra 2 will be the same in a phone, netbook or tablet, surely. So the performance figures should be the same as if it was in a phone. I agree that the battery life details would be fairly useless at this point.

Remember that, while a bigger device can have a bigger battery, it will have a much larger screen - which is where most of the power will probably go.
Posted by drunkenmaster - Tue 31 Aug 2010 15:50
ITs been touted to appear in dozens and dozens of phones and small devices and failed to appear in a single one, largely rumoured due to missing design specs on power. Meaning even though its still getting design wins for smaller devices its being dropped more from the initial design wins that its getting new ones, which may or may not ever appear.

Its being benchmarked, from the sounds of it, in a tablet device, against(I'm honestly not into this mobile everything device malarky I'm a man and can carry a whole 3kg laptop should I want to so no idea on the names of things) what I think are phones mostly?

The fact is things going into phones over tablets will almost certainly be running at lower voltage and clock speeds than the chips themselves are capable of. I mean a 35W dual core running at 2Ghz is the same chip that can use 70W at 3.5Ghz, its just being run at a lower setting.

So as said performance without the power ratings is worthless in mobile devices. If its a device thats got a 100gram heavier battery thats got 3 times the life and they've whacked the clock speed and voltage up due to available power, it should be faster.

The question is, how fast is tegra 2 against similar devices with chips designed for those devices? If it was THAT good and fast and efficient, IE maintained the same power/speed ratio of other chips, then wouldn't it be being widely used today, rather than dropped in hundreds of previous design wins?

The firing of the head of the Tegra team, the moving to Tegra 3, the adding in of a large number of other members into the team and roadmaps to move onto and past Tegra 3 as fast as possible says to me no one really wants Tegra 2, and they think 3 will be a flop/stopgap, and 4/5 redesigned with the aim of actually hitting power targets seems likely.

The likely emergence of certain Tegra 2 devices, almost definately if it only appears in 1/2 devices, is most likely Nvidia paying someone heavily to use their chip to make it appear successful, if it was actually successful AND that good, it would be in dozens of devices already surely?