Windows tablets

PowerVR SGX545 supports DX10.1 (http://www.dvhardware.net/article21698.html), PowerVR Rogue supports DX11.1 (http://www.anandtech.com/show/5364/powervr-series-6-rogue-gpus-released-to-licensing). ARM Mali-T604 supports DX11 (http://www.tomshardware.com/news/ARM-GPU-GPGPU-Mali-T604-Mali-400,11616.html). Tegra 4 supports DX11 (http://www.hi-technonews.com/nvidia-tegra-4-to-8-core-processing-directx-11.html). All the new ARM SOCs that will be released later this year (and onwards) will support either DX10 or DX11.

I am sceptical about the reliability of some of those sites.
 
Hmm, that resolution is standard for all Windows laptops. Do they even make 7-inch or 10-inch displays with those resolutions?

Probably the latter.

How well will W8 scale between different resolutions? Will W8 apps. be resolution-independent

There's still resolution dependence as we aren't dealing with vector-based images however, Microsoft heavily recommends if not outright requires 100%, 140% and 180% versions of images. I expect to see 10.1" 1920x1080 screens in tablets when Windows 8 ships. 7" laptops may get 1600x900, giving us a 260 DPI, which is possible to make today, and may be economical in late 2012.
 
I am sceptical about the reliability of some of those sites.
Then you have to do a google search of your own. There were plenty of hits to choose from in all instances. And I am sure you can find the original press releases if you dig deep enough.
 
Then you have to do a google search of your own. There were plenty of hits to choose from in all instances. And I am sure you can find the original press releases if you dig deep enough.

Well I won't exclude the possibility for Wayne being up to DX11 after all, yet that link of yours obviously has quite a few things confused:

Nvidia Tegra 4, the second silicon Wayne reportedly much stronger than the first version, because the package is not less than 8 ARM processor cores to 64 cores and 32 GPUs are compatible with DirectX 11 + and also supports OpenGL 1.x and 4.x and OpenCL PhysX.

It sounds more like someone is confusing Tegra with Denver again.

99,9% from that what floats around in the internet and has not been officially confirmed by the IHV itself (like in the case of Rogue or T604 f.e.) is just hearsay. If I'd dig in comparable speculative newsblurbs of the past Tegra3 should also have "scalar" ALUs by now and around DX10 compliance.

Apart from that you are IMHO correct that if not all at least the majority of future platforms targetting win8 and beyond will be either DX10 or DX11.
 
Apart from that you are IMHO correct that if not all at least the majority of future platforms targetting win8 and beyond will be either DX10 or DX11.
And DX11 featurelevel 9_3 isn't that bad either. It has MRT support (4 RTs) and all the key features required to program console quality graphics. The new indexed (partial) constant buffer API should work on all devices (not just DX11.1 devices). Tessellation and geometry shaders aren't important, as long as we get the better API (with lower overhead and more strict feature requirements).
 
And DX11 featurelevel 9_3 isn't that bad either. It has MRT support (4 RTs) and all the key features required to program console quality graphics. The new indexed (partial) constant buffer API should work on all devices (not just DX11.1 devices). Tessellation and geometry shaders aren't important, as long as we get the better API (with lower overhead and more strict feature requirements).

Tegra3 should be already DX9L3. It's my gut feeling though that the bigger point of interest for anything =/>DX10 isn't as much what you mention above but rather GPGPU related.
 
Tegra3 should be already DX9L3. It's my gut feeling though that the bigger point of interest for anything =/>DX10 isn't as much what you mention above but rather GPGPU related.
Yes, of course compute shaders are the most important DX11 feature, but sadly the 9_X feature level devices are not going to support them, and neither are all DX10 devices. Most developers aren't going to spend the extra resources to do their whole lighting and post processing pipeline using compute shaders, if they still need to support the old DX10 hardware (and run a separate pixel shader path to them). Compute shaders would be really important to mobile devices, as they allow more efficient implementation of many graphics algorithms. Extra battery life (and performance) is very important, and thus brute forcing with pixel shaders isn't the most efficient way to solve issues (if an alternative compute shader algorithm could run much faster and use less resources).
 
To what extent does using compute shaders help in improving efficiency of your rendering engine?
Compute shaders have access to shared memory (basically a small on chip work memory area), and it can be used to reduce memory traffic and to reuse already calculated data. For example you can share calculations and fetched memory with neighbor pixels (bigger blocks of pixels). With a traditional pixel shader architecture you have to calculate/fetch everything independently for each pixel (and this often causes extra work). You can also implement more sophisticated algorithms with thread block synchronization instructions instead of using simple brute force algorithms that scale badly.
 
Should we expect for linux kernel based platforms (Android, iOS) the primary concentration for GPGPU to be on OpenCL and windows based platforms DX?
 
http://www.engadget.com/2012/04/27/business-minded-hp-slate-8-tablet-surfaces-in-leaked-image/


here is one of the first windows 8 tablet leaks from HP .

10.1 inch sceen
Windows 8 professional os (this means its a x86 chip)
Multi-touch or digital pen

8-10 hours of battery life
9.2mm thick
.68kg (1.5lbs)

This looks very good to me.

The only question is what is the resolution of the screen and if its wacom pen input .

If the windows 8 tablets are all going to be getting battery life like this in this form factor that will be darn good

For comparison

The ipad 3 is
9.4mm thick and 1.44 lbs heavy the 4g verison is 1.46

The ipad 2
8.8mm 1.35 lbs (3g)

Both are listed at 9 to 10 hours per apple.com


On the andriod side

Transformer prime is 8.38 mm thick and weighs 1.29 lbs and up to 12 hours battery life
 
I've been waiting to take the true tablet plunge.
Right now I have a rooted Nook Color with Android Gingerbread. And while it offers a lot of functionality, I've always viewed it as a stop gap true Tablet experience.

From the Windows 8 consumer preview vids I've seen on Youtube, it looks pretty promising.
Ultimately, I've decided to skip the Ipad 3. It will come down to an Android or Windows 8 tablet for me. Pen support functionality from the get go, is a big plus to me.
 
Pen support functionality from the get go, is a big plus to me.

This...is an absolute must for me. And not just pen support, but active digitizer pen support. After having used a touch only slate for over a year now, I have to say a touch only slate is by far...

The most horrible thing in the world for me and my useage patterns. :p

If it doesn't have an active digitizer pen, then it's useless to me. Touch is just so inferior for input purposes for the majority of the things I need a slate for. And capacitive touch pens are just universally horrible.

Never again will I ever buy a slate/tablet/whatever that is touch only.

Regards,
SB
 
Maybe it's Clover Trail-W as speculated in the Intel thread.

mabye , depending on when its released exactly it could be a brazos 2.0 or another amd one.

Thats what i'm looking for a dual core 2ghz apu . Should allow me to play alot of diffrent games and should still get good battery life if its a 32/28nm chip.
 
That HP Tablet is almost certainly based on Clover Trail. Intel's experience in getting Medfield to have competitive battery life in Smartphones are obviously being carried onto Windows 8 devices.

It can't be about CPU alone. If CPU was the sole determining factor, Windows-based Atom devices should have gotten far better battery life. The platform based power management they have been working on with Haswell* and learned from getting Smartphones to get competitive battery life is the reason.

*The 20x reduction in idle power

The summary is basically:
-Get even lower power states for all devices. That includes not just the SoC(CPU/Graphics/Memory controller/IOH) but memory itself, various sensors, and anything that is needed to connect to devices(PCI Express/USB/Legacy).
-Use an OS that supports activity based interrupt(like Windows 8) rather than periodic interrupt(like Windows 7)
-Defer interrupts when possible to enable idle power to be long as possible between wake ups. Longer idle times mean devices can go into deeper sleep modes
-Faster P/C State transitions
-Work with every vendor to make sure they meet the requirements
 
Last edited by a moderator:
There are two reasons I'm wanting a Win8 tablet over the others.
- ability to easily mount network shares and access from any app. Android can do this if you root and get a custom kernel. PITA. IOS doesn't seem able to.
- maybe play some x86 turn-based strategy games or RPGs.

More CPU power will definitely be welcome for PDF/CBR/CBZ.
 
Back
Top