NVIDIA's Ion Platform: Performance Preview
by Anand Lal Shimpi on February 3, 2009 9:01 AM EST- Posted in
- GPUs
Ion’s Performance: Generally Faster
I ran a couple of real world application tests on the Ion to see if it offered any tangible performance benefits outside of gaming. To my surprise, it did.
The first test was our standard Photoshop CS4 benchmark. I don’t expect anyone to want to edit photos on an Atom based machine but if you find yourself in such a predicament let’s see how the Ion performed:
NVIDIA’s Ion platform is 20% faster in our Photoshop benchmark than Intel’s standard 945G.
The next test is far more basic and very realistic: it’s a WinRAR compression test. I’m simply taking several files and compressing them using WinRAR. This is a very CPU intensive task and actually takes an annoyingly long time on any Atom based machine.
Here Ion is 15% faster than Intel’s standard Atom platform and it’s a very realistic usage scenario, even for an Atom.
While you can argue that you may not game on an Atom based machine and you wouldn’t want to edit images in Photoshop on one either, compressing files will happen and it just happens faster on Ion.
The advantage here appears to be memory bandwidth. While 945G is limited to a single DDR2-667 channel, the Ion reference platform runs DDR3-1066 yielding nearly 60% more memory bandwidth.
Facebook
Twitter
RSS
69 Comments
View All Comments
sprockkets - Tuesday, February 3, 2009 - link
Anand reported the GN40 will be mobile only and have no SATA on it. Not the same markets.SirKronan - Tuesday, February 3, 2009 - link
I have a car that has a decent-sized LCD screen built into the front of the dashboard (review camera, climate controls, etc.) From the moment I heard about Ion I got excited. I want to build one of those things into my car. I sure hope they are released with good pricing before the summer's over, because that's what my summer project will be. Kits online abound that add an external video input into my car's LCD. With one of those kits, I want to integrate the small Ion PC into my dashboard. I will also connect bluetooth and wireless N to it. I will integrate a USB port right into the dash as well. Using a solid state disk, I'll be able to consume a minimum amount of power, but it will be connected to everything. I'll have music on it, when we park in our driveway we can wirelessly sync content from our home media server, use a bluetooth mouse on the the dashboard, a bluetooth keyboard when we're at a random hotspot to surf the web. I'm so stoked for this. It will be noticeably faster than the 945 in all tasks, and playback video flawlessly. We won't have to pack around the laptop in the car to play videos for my son anymore. And, the project won't cost an arm and a leg ... hopefully.Thanks, Anand. Good review.
Casper42 - Tuesday, February 3, 2009 - link
And with the 3D Acceleration, you could use a mapping software that does true 3D rendering as opposed to 2D sprites.SirKronan - Wednesday, February 4, 2009 - link
Absolutely. I plan on installing a USB based GPS as well. Will be a lot bigger screen than our current Garmin.chucky2 - Tuesday, February 3, 2009 - link
I was thinking along the same lines: This will be great for the Car PC crowd that wants video in their ride.Chuck
Bull Dog - Tuesday, February 3, 2009 - link
I thought black was ok, not pretty but ok.White on the other hand is purely hideous.
shmina - Tuesday, February 3, 2009 - link
The GeForce 9400M in the MacBook is spec'd at 16SP, with a core running at 450MHz with a 1.1GHz shader clock.The specs you stated are for the desktop GeForce 9400 that has a 30W TDP vs. the 15W for mobile.
deputc26 - Tuesday, February 3, 2009 - link
Wait 20W at idle???? My Laptop with T8100, 8600m GT and 1440x900 screen, 802.11n and bluetooth(the huge power hog) draws only 11w at idle. How is this low power??7Enigma - Tuesday, February 3, 2009 - link
I'm interested in the explaination here as well...SilentSin - Tuesday, February 3, 2009 - link
I noticed this too. Was powerplay not fully implemented in this thing? The Atom 330 is spec'd for 8W TDP vs. 2.5W for the 230 used in the Eee Box so the difference in power usage here seems to reflect that exact bump of about 5.5W in the CPU. So that means those power consumption figures point to the NV chipset using just as much power as the dinosaur 130nm 945G chip. What gives? I know there are a metric crapton of features added by using the NV chipset, but I would still expect an overall reduction. Remember that most reviews of the Atom platform seemed to blame the 945G for the "high" power usage figures. I would have hoped a more focused mobile-oriented chipset would have done better. Maybe the problem is in the demo box? It looks as if they are using dual phase power distribution where they could probably get away with single phase for something like this, maybe that is the culprit.