http://www.pocketables.net/2009/04/samsung-mondi-wimax-mid-powered-by-arm11-cpu.html

I’m rather sure that the performance is not 4 times faster between a ARMv7 & ARMv6 design. At best your looking at twice the performance, with only on a few subsystem on the CPU design with a greater gain ( but those do not add to the general speed as to create a 4 times speed increase ).

There has been some significant discussion going on about that one, on the OpenPandora forums, and you can look up the data in the threads there.

But having a browser loading a page 10* faster, that will be more up to:

- Faster Memory access ( higher speed memory )
- Faster internet access ( 3G ) vs whatever in the past.
- Faster Storage ( SD card speeds have been going up plenty )
- Faster browsers
…

Do not expect the DSP to add that much to the general speed. The DSP / Neos is like SSE on the PCs. Sure, it gives a boost, but it does not magically turn something into a speed beast. And those bring there own set of problems into the equation.

It all depends on how they compared the N810 vs the Mondi… Its like comparing apples vs oranges… Its like: Look, with the screen of, nothing running, our laptop does 6hour. Where as the competition with the screen on, running xxx threads, is only doing 1 hour.

Take it from somebody who has falling before into that trap, lets first some real world benchmarks, before running into the claims with PR written all over it

I’d love to see a device that delivers ten times the “processing” power with x hours, but lets first stay realistic… Dreams can quickly turn into nightmares

The tegra has only one cpu. You are mixing the fact that the tegra has the CPU & GPU on the same die. Thats just like the TI OMAP, and the Snapdragon…

Also, it has been pointed out that the limitation on the Tegra is going to be its older cpu core. A ARM11 is in reality a older ARMv6 ( confusing is it not ), while the TI OMAP, and Snapdragon use a Cortex A8 ( ARMv7 ).

Like i posted above, a ARMv6 has a performance off around 1.2 / Mhz, while a A8 ( ARMv7 ) can do about 2 / Mhz.

There is a Multi Cpu core in design for Arm, but don’t expect to see that in production for the next year or so.

But if you’re connected to the internet all day to run the skype, how about the battery life?

2 or 3hrs is too short to be a phone~

This is shaping up to be my “most likely to replace my iPaq” device. Either this or the Toshiba TG01.

Can’t wait.

S

I mentioned this in a previous article; I suspect this device has been commissioned by Clearwire as a WiMax showcase device.

Here’s a question though, if you’re able to connect to the internet and run Skype for 24hours with Skype-in and Skype-out services, does it make this a mobile phone?

Also there’s the current trend of many companies putting ARMv6 chips in their summer 09 phones and delaying the ARMv7 tech to Q409(Toshiba and Palm being exceptions). I do hope I’m wrong though-the Mondi would be a great Snapdragon device!

I think we can safely say that the end user would see a significant improvement and that’s already been proven with the Archos 5.

But a claim of 10 * faster then a N810… Even if it includes a TI OMAP 3630 ( 600Mhz ), thats just about what? 3? times faster…

The only way i can see a claim like that, if its running a Snapdragon @ 1Ghz, vs the n810 with no hardware drivers for its gpu. And they benched on that, or something…

Problem is, that article claims processing power, and that indicates CPU. So unless they included a 1.5Ghz Snapdragon, i don’t see that happening.

Lets do that math. Lets assume the ARM11 does 1.2 vs 2.0 Cortex A8. At 400Mhz, that give a score of 480. For 10 times the processing power, the Cortex needs to be doing 2400Mhz. Fat chance for that one. Unless they also include the DSPs in some way, in there calculations… Now, lets say, 6.5 times faster, then in theory, a Snapdragon at 1.5Ghz can do it.