Anandtech: NVIDIA's Ion Platform (Atom+GeForce 9400M)

[dougz]

Jon Stokes compares ION to AMD's Yukon & mobile Radeon, as well as VIA's Trinity platform. Claims ION should draw about 18 watts.

http://arstechnica.com/news.ars/post/20 ... -2009.html

[|Romeo|]

Jon Stokes compares ION to AMD's Yukon & mobile Radeon, as well as VIA's Trinity platform. Claims ION should draw about 18 watts.

http://arstechnica.com/news.ars/post/20 ... -2009.html

18W?!? That is going to be really toasty. And also sadly makes it completely useless to me

Back to waiting for Moorestown/Mobile ITX I guess

[Ant6n]

toasty?

... I have some mobile dual core run overclocked and passive, and the thing has a tdp of more than 20 Watts - and that's just the cpu, without the chipset.

My laptop has a pentium m 1.7 GHz, and at the lowest power setting it runs completely passive, at a tdp of 10 Watts - again without counting any chipset, hdd etc.

[dougz]

18W?!? That is going to be really toasty. And also sadly makes it completely useless to me

I think it all depends on the intended usage. ION is clearly not targeted for low power Mobile Internet Devices.

OTOH, HD TV at 18 watts (plus storage devices) is pretty good!

[|Romeo|]

In my oppinion, based on experience with a Via Pico ITX and the size of heatsink shown in the pictures of the Ion, it is going to get hot.

No doubt time will tell if I am correct or not, but I'm still dissapointed by the high power consumption.

[dougz]

If you are able to hang tight for a couple of months, do so. If you missed the news about nVidia's Ion platform-based netbooks, affordable machines as tiny as the Aspire One capable of playing 3D games and encoding video could be just over the horizon. The technology may add a little more to the price ($50 max, according to nVidia).

Source -- http://tech.yahoo.com/news/pcworld/2008 ... salesdeals

Emphasis added.

[dougz]

VLC Media Player To Receive VDPAU Support

Posted by Michael Larabel on December 23, 2008

Back when NVIDIA introduced VDPAU they had provided a set of patches that implemented this video API within the MPlayer and FFmpeg projects.

Since the introduction of the Video Decode and Presentation API for Unix, the community has come up with VDPAU support for MythTV and most recently the Xine media player picked up VDPAU support. This afternoon we now have word that the VLC media player will be picking up VDPAU support.

The VLC project currently has two developers working on implementing this video API that currently is only used by NVIDIA's binary driver. The confirmation of this work can be found in the VideoLAN Forums. Implementing VDPAU will allow for HD video playback on Linux even with very low-end hardware.

Source -- http://www.phoronix.com/scan.php?page=n ... &px=Njk1NA

[MiKeLezZ]

18W?!? That is going to be really toasty.

LOL!

[mentawl]

Intel insists Atom only for Intel chipsets, does not want to share nVidia's Ion

Anyone else find it supremely amusing that nVidia is upset about this, when one considers nVidia's similar desire to not share SLi functionality with other core logic chipmakers? Maybe Intel can trade Atom for a general unlocking of SLi, but that's probably just wishful thinking

[frank2003]

So this Ion thing is a unilateral agreement? I guess nvidia jumped the gun and p*ssed off the giant.

[line]

Nvidia already has permission to play in Intel's desktop and mobile markets. I'll withhold sympathy for Nvidia's Ion woes until the G45(M) vs. GF 9300 scene heats up...

[frank2003]

Intel insists Atom only for Intel chipsets, does not want to share nVidia's Ion

Update:
"There is nothing preventing vendors from using the Ion platform. We sell Atom as a stand-alone processor, or as package with chipset," said Bill Calder, an Intel spokesman, in an e-mail sent to InternetNews.com. [Source here]

So we should see some Ion desktops (and hopefully motherboards) in March and netbooks in June. [Source here]

[derekva]

What front end are you guys planning on using with this little box?

Boxee and XBMC don't support DxVA acceleration. Those are both excellent front ends but this platform would be nearly worthless with that software. Or at least not any better than a Xbox. SageTV would be a good choice and maybe even VMC. Linux now has hardware acceleration for h.264 with Nvidia cards also.

Seems like this should work fine with Windows 7 Media Center (once Windows 7 ships).

-D

[zapwizard]

If you take a look at the newest Mac Minis they are essentially the nVidia Ion platform with a faster Atom on-board.

They claim 13-watts at idle.

I wish you could buy the reference design they sent to all the review websites. It would be great to see just how far you can push one.

[jessekopelman]

If you take a look at the newest Mac Minis they are essentially the nVidia Ion platform with a faster Atom on-board.

They claim 13-watts at idle.

I wish you could buy the reference design they sent to all the review websites. It would be great to see just how far you can push one.

No. The Mac Minis have a Mobile Core 2 Duo. That is not any sort of Atom.

[dougz]

The problem is that while the Ion’s hardware decode engine can handle the video decode just fine, there’s a lot the CPU still has to do. When playing a standard Blu-ray disc there’s a lot of decryption that the CPU must do in addition to simply streaming the data to the GPU. On a modern day processor this isn’t a problem; on the Atom, it’s a serious workout.

A dual-core Atom is, however, capable of playing back all Blu-ray titles as far as I can tell; at least the most stressful ones out today....

Final Words

NVIDIA’s GeForce 9400M chipset alone isn’t enough to ensure problem-free Blu-ray playback with the Intel Atom processor. The problem is, as I’ve said before, that NVIDIA’s Ion isn’t a balanced platform. It uses a GPU that’s much too fast for the CPU it’s mated with.

NVIDIA wants the world to believe that the GPU is all that matters going forward. While the GPU is a very important part of many systems today, it does not eliminate the need for a fast CPU.

If you want an Ion system to serve as your HTPC or if you just want to watch Blu-rays on your Ion netbook, you’ll need a minimum of a dual-core Atom 330.

Ion systems are expected to begin shipping in the next few months.

http://www.anandtech.com/video/showdoc.aspx?i=3526&p=2

[Scoop]

The problem is that while the Ion’s hardware decode engine can handle the video decode just fine, there’s a lot the CPU still has to do. When playing a standard Blu-ray disc there’s a lot of decryption that the CPU must do in addition to simply streaming the data to the GPU. On a modern day processor this isn’t a problem; on the Atom, it’s a serious workout.

A dual-core Atom is, however, capable of playing back all Blu-ray titles as far as I can tell; at least the most stressful ones out today....

Final Words

NVIDIA’s GeForce 9400M chipset alone isn’t enough to ensure problem-free Blu-ray playback with the Intel Atom processor. The problem is, as I’ve said before, that NVIDIA’s Ion isn’t a balanced platform. It uses a GPU that’s much too fast for the CPU it’s mated with.

NVIDIA wants the world to believe that the GPU is all that matters going forward. While the GPU is a very important part of many systems today, it does not eliminate the need for a fast CPU.

If you want an Ion system to serve as your HTPC or if you just want to watch Blu-rays on your Ion netbook, you’ll need a minimum of a dual-core Atom 330.

Ion systems are expected to begin shipping in the next few months.

http://www.anandtech.com/video/showdoc.aspx?i=3526&p=2

The article has been pulled. Anand got a note from Nvidia saying the test data wasn't correct and want to test it for themselves. Anand being a good boy took the article offline.

[dougz]

The article has been pulled. Anand got a note from Nvidia saying the test data wasn't correct and want to test it for themselves. Anand being a good boy took the article offline.

In the comment thread, Anand says

RE: Will we ever see the original article again? by Anand Lal Shimpi, 2 hours ago
I won't publish NVIDIA's findings, only my own. The thinking is that there was an issue with the version of PowerDVD I was using which has since been corrected. I won't say for sure until I re-run the tests myself but I have reason to believe that NV is correct in this case. I simply didn't want the story to spread like wildfire if there was any chance that my findings were incorrect. I've always been in this for the truth, not the sensation

http://www.anandtech.com/weblog/showpost.aspx?i=581

Anand did the right thing to pull it, pending resolution of the issue. One poster in the thread says

I don't doubt ION was a bust given powerdvd was involved by mikesm, 3 hours ago
When are AMD and Nvidia going to learn that they can't depend on the bozos at Cyberlink to write code that takes full advantage of their hardware for playback? Anyone who has built an HTPC knows how disappointing the support for hardware offload is in PowerDVD and how it varies from release to release.

Most of the time the CPU can pull the chestnuts out of the fire, but with ION it just plain doesn't work unless the GPU is doing all it can.

but poster mikesm cites no evidence. http://www.anandtech.com/weblog/showpos ... 2#comments

It will be interesting to see how this nets out.