Slashdot Mirror
nVidia Preview 'Tegra' MID Platform
wild_berry writes "nVidia have previewed their Mobile Internet Device platform which will be officially unveiled at Computex in the next few days. The platform features CPU's named Tegra paired with nVidia chipset and graphics technology. Tegra is a system-on-a-chip featuring an ARM 11 core and nVidia's graphics technologies permitting 1080p HiDef television decode and OpenGL ES 2.0 3D graphics. Engadget's page has more details, such as the low expected price ($199-249), huge battery life (up to 130 hours audio/30 hours HD video) and enough graphics power to render Quake3 anti-aliased at 40FPS."
But seriously, this sounds interesting. If they actually manage to pull it off, this might actually make TV on the go a real possibility (compared to strain your neck trying to watch Sex and the City on your phone...).
Now the only question is, how heavy is the battery to allow for such a long lasting device. You can't tell me it actually is this efficient, if it boasts that kind of computational power.
Somehow, I doubt it though.
The simple truth is that interstellar distances will not fit into the human imagination
- Douglas Adams
I've been waiting for ARM laptop thing. Real battery life! Why do I need x86 compatibility? Give me battery life every time.
I almost bought an Asus EEE pc this weekend, this is worth waiting to see how it is implemented in consumer devices. Give me a small laptop type that can run linux and I'll buy one or two. Heck, 30 or 40 hours would be enough battery time, don't need 100.
Kindness is the language which the deaf can hear and the blind can see. - Mark Twain
Too bad this comes from nVidia, who is hostile to open source. If it were someone else we might at least hope to get specs. Oh, well...
When I heard that a company was making an inexpensive computer with great battery life, adequete performance and it was going to be 'ultra portable', I was so happy! ... then they released it and it was more expensive than originally planned, and not quite as robust.
... and now nVidia is going to do the same thing to me.
If it can run ffdshow or VLC at 1080p then we're talking something special.
I read an article about the Atom platform, which competes in this space. Apparently only the most powerful version of Atom would have enough oomph to run Vista, so can this nVidia MID handle it acceptably? (I know, the review mentioned it runs Windows Mobile, but I'm curious.)
I'd certainly be willing to offer my meager talents to the effort for THAT kind of battery life. Will an ITIL metrics slide help? :/
Invenio via vel creo
iPhone 3.0. Actually the current iPhone uses Power VR MBX and the new one is rumored to be using the Power VR SGX graphics. The Power VR VXD video IP core can supposedly "supports 1080p H.264 Main/High Profile decoding, as well as VC-1 and a variety of other standards" http://www.beyond3d.com/content/news/638 http://www.appleinsider.com/articles/08/04/30/apples_bionic_arm_to_muscle_advanced_gaming_graphics_into_iphones.html
Patience...
Pandora comes...and it is looking like it's going to largely deliver on the "promises" it makes.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
May it run Doom instead?
Sounds like an interesting toy, but aren't we twisting the measurements a bit here? Quake 3 came out in 1999. Any modern graphic chip has the graphics power to render Q4 at much faster than 40 FPS. Of course, there's the important question of "do you have the computing power behind the graphics power to make the game playable without lag or stutter on anything but a non-trivial map?", as is "do you have the system resources to get a new map started and get into the game before the other players all have multiple frags ahead of you?". And perhaps the most important question is "at what resolution?". Talking about playing a game anti-aliased at 40FPS but not saying what resolution you are playing at is completely meaningless. While this hardware may be able to 1080p HiDef video, there is an awful good chance that that lame benchmark "spec" is based on a much lower resolution.
I sure hope that this doesn't lead to further hype and dumbing down of video specs. Look for new graphic chips that can run Wolfenstein at 1692 fps or Pong at 31500 fps anti-aliased.
I'm an American. I love this country and the freedoms that we used to have.
At best, maybe they'll put some hooks in ffmpeg's library (or directly in VLC as an alternate engine) to call their BLOB to handle the accelerated decoding.
At worst you'll have to use a binary only nVidia-specific player. And given that the ARM+nVidia platform isn't going to be very popular fact, probably not a lot people are going to reverse engineer it (ala "Nouveau" project) - expect if their ARM stack is exactly the same as their x86 and thus the work could harness what's already been done by Nouveau.
But the whole thing is going to be opaque.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Does it run linux, does it blend, in Soviet Russia a beowulf cluster of their new overlords welcomes ME!!!!
The PowerVR vs. nVidia comparison is approximately the same as the ARM vs. Intel Atom.
nVidia are producing classical graphic cores.
PowerVR are employing specific techniques (Tile-Based Deferred rendering) which enable them to cram the same performance using a lot less transistors and running at lower clocks.
The nVidia SoC is probably more targeted toward sub-notebooks, big multimedia PDAs (As a example, the TapWave Zodiac was based on an ARM and an ATI Imageon running PalmOS 5) and small internet-enabled appliances.
Smart Phone will probably use whatever is less power hungry and go for PowerVR's designs.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
for those who actually enjoy RTFA'ing and want a bit more comprehensive info than a BBC fluff piece, nvidia's marketing page, and some pretty vids on engadget:
http://www.tgdaily.com/content/view/37729/135
The APX 2500 is far more interesting to me than the 600/650. Qualcomm and Broadcom better watch their backs.
Protector of Capitalist views,
Meorah
Well, the difference is that they don't use the mobile phone as a PC. It's an appliance. Same as a fridge, or a DVD player, or a TV, or their fixed landline phone. If it does its job, why would you care if your fridge has an x86 in it? Most "normal" people I know don't really do much more than phone on their mobiles and sell the occasional SMS. Very few even realize that they could run any other program on those, much less actually download one, so compatibility doesn't play a role. But I dare say that with a laptop, the expectations are a bit different. People start having these ideas like "can I open the .xls file my boss sent me?" or "can I edit photos on it?" Right before, "ugh, why can't I just use Excel, which I already know how to use?" or respectively, "whoever designed the interface of Gimp should be anal raped with a porcupine. Why can't I use my usual editor?" And that's when compatibility starts to matter. Not by itself, but by having access to the same software.
A polar bear is a cartesian bear after a coordinate transform.
I will definitely be getting one of these tegra powered devices.
Hopefully, a phone. And, hopefully it won't cost $400.
really, a smart phone with that chipset should only cost about $200.
with 1080p tv/video and gaming.
They're using their grammar skills there.
Does it?
Except Pandora promises at best a third the battery life. Then again, Pandora is due out very soon, and reading about both it sounds like Pandora is the type of machine nVidia would expect Tegra to be used in.
Centralization breaks the internet.
It's NVIDIA now, not nVidia. It's been that way for a year or two.
I don't need mobile TV. What I need is a few cheap, reliable, fanless, low power media terminals to stream HD video date from my Gbps LAN server, convert it into 1080p HDMI/DVI for my big TVs.
So what I need is some Tegra PCs with minimal HW (maybe a DVD/Blu-Ray player, but no floppy, modem, or really even a HD - just 8GB Flash and PXE boot) that's mainly LAN and HDMI/DVI connections, running Linux, and full-featured Linux drivers. Preferably open-source drivers that we can tweak to work right, but which get full performance from the HW.
--
make install -not war
If this thing isn't going to run Linux, then I don't know what it is going to run. Certainly not the 'nearly free', stripped down Windows XP for cheap portables.
nVidia would be pretty silly to build this thing and not to provide a proper driver for the only OS it'll probably work under. Of course, if this thing takes off, Microsoft probably will come out with a 'mini XP for ARM-based cheap portables'. But nVidia's got to feed the Linux chicken in order to lay that particular egg...
Posted from my Android phone. Oh, I can change this? There, that's better...
Sixteen years ago there was a first ARM laptop. Running Risc Os but without any doubt some could boot with BSD if the owners were masochists.
Perhaps this technology could be used to produce a very small quiet and low power consuming mythtv box...The noise of my current system can be annoying when trying to watch a movie, but i didn't want to skimp on the cpu because i wanted to play 1080p video on it.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
I'm pretty much sure that you can't play Crysis on a Tiler, unless you make the tiler so much over-complicated that it looses its advantage over classical architecture...
BUT, common, I'm speaking about *SmartPhone*. Nobody's going to play Crysis or anything that has more than a couple of kilobytes worth of shader code on a 320x240 resolution that fits in you pocket.
Besides, this kind of situation isn't very likely to happen any way because :
- Even on the desktop you won't encounter lots of kilobyte long shader for the simple reason that there are currently a lot of video boards on current gamers' machine that won't be able to handle such long shaders and will abort execution or even fail to compile them.
Even today, when both major player have already 2 generations of DX10-compatible hardware (GeFroce 8x00/9x00 and Radeon HD2x00 and HD3x00) and there's even a third one comming into the pipeline (HD4x00 will hit market soon), even PowerVR has a DX10.1 compatible chip. But lots of players are still having GeForce 6x00 or 7x00 or Radeon X series in their computer. Thus most game only have optional DX10 support or even separate patches.
- Also the main caracteristic that enable TBR is that chip have huge onchip memory (for tiles and for texture cache). This already takes several kilobyte of on chip memory. A couple of kilobytes more to cache the shaders isn't an issue.
Current limitations on shader for this chips are more due to the chip being lighter to have lower thermal and power limitation, and lack of usefulness on the small screen than technical limitations. I'm not sure where you think they're getting away with less transistors for a tiler either, there's a whole binning engine (and the associated bandwidth) which a classical architecture doesn't have. This idea is based on comparative transistor count back at the time of the Kyros. Of course since then, on big graphic card, the shader complexity has drastically increased. But then again, as I've said, I'm talking about phones. Nobody is going to put a GeForce 9800 inside a phone. Nobody except Intel's Larrabee engineers are thinking about running GPGPU on a smartphone.
By the way, binning can partly be done by the driver, and doesn't necessarily need expensive hardware.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
In a perfect world this might be interesting. In the real world, if you build such a platform, I can assure you that some script kiddy is going to play games with your system that you will not like.
I'm an American. I love this country and the freedoms that we used to have.
It's amazing- so much innovation for such a small amount of money... and here St. Nic and the OLPC crew are trying to pedal tech that was outdated at least ten years ago.
But as everyone, even people who used to work for OLPC, keep saying... OLPC was never meant to succeed. It was always just another high profile scam, run by hardcore conservatives, and meant to steal money from people and countries who don't have any money to begin with... and are thus less likely to seek redress for being defrauded. "Greed is Good", the conservative mantra.
I which case, we actually both agree, given that a couple of posts ago I mentioned that this new chip will be perfect for sub-notebooks and PDA. This even makes more sense if the later is coupled with one of those laser-based embed projector. The the hidef resolution will definitely make sense. even if you disavow the use of it, i fail to understand how anyone could claim the lack of the feature is a good thing. I'm not saying that it's a good thing that these features are absent. I'm saying that ultra-long shaders aren't currently widespread. At all. And I don't really see them starting to get massively used in every possible device to the point than even the current-day phones (like the iPhone or BlackBerry) need to start including them ASAP.
On the other hand bigger multimedia device (the PDA that try to both steal the lunch from the iPhone *AND* from the workstation) would definitely have a potential use for multimedia acceleration (currently both ATI and nVidia are working on GPGPU-accelerated multimedia decoding).
I'm sorry I misunderstood you, I was initially picturing someone trying to run Folding@home on their latest phone/mp3player/pda hybrid. I think we definitely were speaking about different class of devices. it requires more advanced caching / buffering, but that should not be a dealbreaker. especially when we start loading our chips with massive onboard caches -- a secret well loved by the gamecube for example. And which is the sine qua non requirement for tiler architecture, and was also used on those kind of architecture for a long time too.
And was also feature on the PS2 which had 4MB of embed ram (big amount back then) inside its Graphics Synthesizer.
You initial comment was that big shaders suck on tiler architecture because the tiler has to load all of them at the same time. Then now you report that bigger onboard RAM are a solution for bigger shader.
I was already saying the tiler architecture requires the chip to have a big amount of on board fast memory and a couple of kilobyte used by shaders isn't that much critical. So the argument of long shaders doesn't stand against tilers.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]