Windows tablets

[Deleted member 13524]

Well, changing the screen orientation in windows through the display manager also takes about 5 seconds.

I guess that was never something MS really thought about. But it should be fixable with an update of some kind.

 

[eastmen]

Here is a list of tablets coming out the rest of the year from what we know. There are a few interesting ones

http://www.tabletpcreview.com/defau...e+schedule+android+honeycomb+webos+blackberry

and the other will have an AMD Brazos processor plus a display-enhancing chipset. Both tablets are supposed to have a battery life of approximately six hours and will come equipped with an accelerometer, light sensor, and Wi-Fi.

wonder what the display enhancing chipset is

The lenvo is also interesting.

 

[Laurent06]

Last time i checked windows enables scaning of hardrives for its search function and that is constantly on. Disabling it on my laptop can give me another 20-30 minutes of battery life

I was talking about the kernel itself, not the services...

 

[tangey]

apologies if this one has already been posted, but the following slate/tablet seems a nice twist that answers some of the tablet issues. magnetic rigid cover that doubles as a wireless keyboard. The video shows it in operation. 3G, GPS, HDMI output, Oak trail runninf either windows or android.

http://www.itechnews.net/2011/04/18/evolve-iii-maestro-s-slate-pc-powered-by-oak-trail-video/

 

[Simon F]

http://www.itechnews.net/2011/04/18/evolve-iii-maestro-s-slate-pc-powered-by-oak-trail-video/

Interesting. Bit confused, though, by the written report saying it was "starting at AUD$799" yet in the video it gives a lower price in USD, yet the Australian dollar is worth 7% more than its US counterpart.

 

[Malo]

Interesting. Bit confused, though, by the written report saying it was "starting at AUD$799" yet in the video it gives a lower price in USD, yet the Australian dollar is worth 7% more than its US counterpart.

Price differences between Australia and US never really update. Distributors in Australia basically base everything off dollar differences from years ago. It got to the point where it was considerably cheaper buying equipment overseas and shipping it.

 

[AlphaWolf]

That effect is not limited to Australia.

 

[swaaye]

Well, changing the screen orientation in windows through the display manager also takes about 5 seconds.

Yeah on my old EeePC with GMA 900 it takes 3 seconds with a Ctrl-Alt-arrow key. Android does it almost instantaneously. I wouldn't say that a few second delay is a big problem though...

 

[eastmen]

http://www.engadget.com/2011/04/27/windows-8-settings-page-suggests-tablet-compatibility-embraces/

Engadget believes in a q1 2012 release date for windows8

they had a video of the metro ui up but ms had it pulled from youtube

http://www.youtube.com/watch?v=t90YO6PcU1k

here is one

 

[Exophase]

acer w500 with the c-50 will get 4-4:30 minutes of 1080p playback as tested by users and reviewers

http://www.techradar.com/reviews/pc.../acer-iconia-tab-w500-936986/review?artc_pg=2

Its not bad , more than double my dm3z with dual core neo when playing back hd .

Video playback isn't supposed to load down the CPU on a platform like this, although it doesn't look like 1080p playback on Ontario is that CPU light for whatever reason.

It'd be interesting to see what the battery life is like playing a game that taxes CPU 100% on at least one core and with GPU pushed up pretty far. Unfortunately all any of the reviews seem to care about is battery life for playing back videos and web browsing or other "light usage." If this is all you want why bother with an x86 tablet to begin with?

 

[eastmen]

its two 1ghz cores i believe cpu load is 70% when playing back video so its going to be pretty much what you get playing a demanding game

 

[Deleted member 13524]

its two 1ghz cores i believe cpu load is 70% when playing back video so its going to be pretty much what you get playing a demanding game

With a 2.5GHz Athlon II X4, that would translate into ~12% cpu load. Using something like a 3.2GHz Nehalem or Sandybridge, we'd fall into ~5% cpu load?

That isn't too far off what we've seen in video acceleration benchmarks.
The UVD 2 has always been a video accelerator, not an independent video processor.

Next-gen Fusion might bring a full fixed-function video processor that should bring down the power demands for 1080p stream decoding.

 

[Exophase]

its two 1ghz cores i believe cpu load is 70% when playing back video so its going to be pretty much what you get playing a demanding game

Well threaded games will be a lot closer to 100% most of the time, and they'll be exercising the GPU. You'd probably get much lower battery life.

The UVD 2 has always been a video accelerator, not an independent video processor.

Next-gen Fusion might bring a full fixed-function video processor that should bring down the power demands for 1080p stream decoding.

This isn't what AMD says, which is that UVD 2 is "full bitstream." Just because these can't run videos autonomously doesn't mean that they aren't doing all the relevant heavy lifting.

 

[Zaphod]

its two 1ghz cores i believe cpu load is 70% when playing back video so its going to be pretty much what you get playing a demanding game

Do the most recent AMD GPUs accelerate deblocking using the UVD hardware? IIRC, up to the HD 4xxx they didn't, and one cold reduce the CPU impact quite a lot by disabling deblocking in the decoder settings. I tested this some time back by clocking an old Athlon XP down to way below 2GHz with an AGP HD4650.

I doubt one would perceive much quality difference on a pad-sized screen, and it might improve battery autonomy during playback notably. Who knows what kind of marginal, yet processing intensive, quality shortcuts an Ipad or Android device does when decoding HD video.

 

[eastmen]

i have no idea about the deblocking

 

[Deleted member 13524]

There's hardware deblocking for H.264 ever since UVD 1.

Either the media player is using it or not, that's a whole other story.

 

[Exophase]

It has deblocking, but it doesn't have the post-processing stuff like denoise and scaling.. but allegedly this stuff is done on the shaders so it still shouldn't take CPU. There's really no reason for CPU utilization to be 70+%.

The real question is if the 70% is really happening on all of these video tests. I wouldn't be so sure.

 

[Zaphod]

There's hardware deblocking for H.264 ever since UVD 1.

Thanks. Just re-tested (Cat 11.2 and latest MPC with default settings, WMR9 renderless) on my old legacy box, and 1080P H.264 Big Buck Bunny hardly made a dent on my ancient Barton running it at 1.4GHz/133.

The real question is if the 70% is really happening on all of these video tests. I wouldn't be so sure.

Exactly. Thus, I surmise that the 70% CPU result quoted above is highly software dependent and that a machine with a C-50 can probably do better.

But then again, that's part of the problem with a Windows tablet in the first place. Software flexibility comes with a cost in terms of integration and optimization that the end user of a more closed system rarely will have to worry about.

 

[swaaye]

I tried a 720p H.264 + 5.1 AAC video on my old EeePC 900 (Celeron M 900 MHz) for kicks again and found that FFMPEG has advanced considerably in the past couple of years because it played fine. I was really surprised. Back when I got the Eee in 2008 it could barely play 480p through FFDSHOW/FFMPEG. 480p H.264 + 5.1 AAC videos are now using only about 40% CPU.

My Nook Color can play 720p as well when using MoboPlayer. I think you need to be overclocked to around 1.1 GHz but it seems that everyone can reach at least that speed (I'm at 1.3 GHz). FFMPEG is getting pretty optimized for ARM these days too.

What's most unfortunate about the OMAP devices is the closed nature of the discrete DSP and its codecs. It's all closed source. Apparently the DSPs vary largely (or only) by clock speed and so a little overclock could make just about any of these DSPs play HD video. But the HD codecs aren't available if your device isn't officially HD capable. Nevermind that the codecs in general could use some smoothing out because they are pretty picky about what they will play. But pickyness has been part of the game with AMD, NV and Intel DXVA acceleration too.

 

[eastmen]

http://www.youtube.com/watch?v=EgDbxMUCpI0

this would have been great for a ms tablet. Wonder if we will see it in windows 8