Intel Atom vs.VIA Nano @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,634
Intel Atom vs.VIA Nano - A battle of the most powerful processors on the face of the Earth? Nope. Who sucks the least? Not really, but closer. Marketing teams will have you believe one thing, but we are here today to show HardOCP readers what they likely care about. Which one just works better, Atom or Nano?

Some of you may be surprised by the results, while others may not. I have to admit, I expected the VIA Nano to come out on top given its superscalar out of order architecture, but I did not expect the Intel Atom to take such an overall beating. The Nano looks like a power efficient version of the Pentium 4 while the Atom looks like a power efficient version of the VIA C7. Everything old is new again?
Click to expand...

Please Digg to Share!
 
Wow, very nice to article to see. When they become mass available VIA Nano based UC's might finally provide a viable (ok that was bad) replacement to my aging Transmeta Efficeon based Sharp given their cost should be lower than the competing Atoms.
 
*sigh

Why does almost no tech site do enough research to understand the rationale behind the Atom?

It's designed to do less work per clock in order to save power (yes that's actually it). It does have hyperthreading to help make up for that but it has to be clocked higher.

If you're comparing a higher clock chip to the Atom, of course, the Atom is going to lose. It's a pointless comparison unless you find an Atom that draws the same power (and thus clocked higher).



The other problem is comparing the Total System Draws and then saying the Nano has lower idle draw. The chipset makes a pretty large difference when the CPU TDP of an Atom is 4W (notice that the difference between Idle and Load for the Atom is about 5W even from the wall). At the moment, the Atom is coupled with a much poorer chipset whereas the Nano was coupled with a purpose driven chipset.

Simply looking at the difference between Idle and Load for the Nano (16W) shows how it's not even the same class as the Atom. It's like comparing a QX9775 (150W TDP) on an Nvidia desktop chipset to a P8600 (25W TDP) on an Intel mobil chipset.

It's fine to compare the chips to show relative performance, but then you get comments in the conclusion like: "When it comes to a sheer performance comparison, the VIA Nano rules and the Intel Atom drools." which not only isn't very good in the journalistic sense, but misses out on the point and an opportunity to explain (that is, give the readers information) that the Atom and Nano have different niches without sensationalism.
 
It was a real world review. They were measuring power from the outlet, which is what I'm most concerned with, so they get a plus from me.
 
That's the thing, even if it's real-world, chipsets change and change all the time. Therefore, you need to give both the full system draw of a current system and an analysis of it. That is, an estimation of what the CPU itself is drawing.

We already know that Intel has chipsets that draw little power. The 945 chipset is really old and could really benefit from things like die shrinks and improved gating technology.

This really highlights why the Atom isn't really a good platform yet. The chipset, the CPU is married to simply draws too much power.
 
ChronoReverse said:
*sigh

Why does almost no tech site do enough research to understand the rationale behind the Atom?

It's designed to do less work per clock in order to save power (yes that's actually it). It does have hyperthreading to help make up for that but it has to be clocked higher.

If you're comparing a higher clock chip to the Atom, of course, the Atom is going to lose. It's a pointless comparison unless you find an Atom that draws the same power (and thus clocked higher).



The other problem is comparing the Total System Draws and then saying the Nano has lower idle draw. The chipset makes a pretty large difference when the CPU TDP of an Atom is 4W (notice that the difference between Idle and Load for the Atom is about 5W even from the wall). At the moment, the Atom is coupled with a much poorer chipset whereas the Nano was coupled with a purpose driven chipset.

Simply looking at the difference between Idle and Load for the Nano (16W) shows how it's not even the same class as the Atom. It's like comparing a QX9775 (150W TDP) on an Nvidia desktop chipset to a P8600 (25W TDP) on an Intel mobil chipset.

It's fine to compare the chips to show relative performance, but then you get comments in the conclusion like: "When it comes to a sheer performance comparison, the VIA Nano rules and the Intel Atom drools." which not only isn't very good in the journalistic sense, but misses out on the point and an opportunity to explain (that is, give the readers information) that the Atom and Nano have different niches without sensationalism.
Click to expand...

I think you missed the point of the article, which states more or less exactly what you said, that the VIA Nano and the Intel Atom are aimed at different markets and shouldn't be directly compared. That was repeated numerous times throughout the article.
 
ChronoReverse said:
*sigh

Why does almost no tech site do enough research to understand the rationale behind the Atom?
Click to expand...


I think that is EXACTLY what we just did. Our own research. Our opinions jive with Intel's mantra. RTFA baby.....rtfa.
 
Hardocp says:

Ey they aint suited for laptops, have i seen a atom in action, no, but through a movie tho, and it is sloooow.

I'm feeling angry by using vista on a 3.6 GHZ QUAD, and im sitting on 3.2 @ home in vista, and im still freaaaaakin pissed off at the slowness.
(Vista)

Linux on a laptop with nano would be great, a atom is just hilarious, its like jumping 8 years back in time.

8 years......

Sure it packs power for 4 watt to be, but they dont have any chipset to back anything up.

amd is comming with their cpu @ 6 watt, BUT! 780G makes up for the 2 extra watts even thou the cpu is indeed slower, who knows, not me.

Videocard is underrated.
VERY underrated, IM NEVER EVER EVER EVER BUYING FULL INTEL BASED LAPTOP!

unless they are at amd/nvidia level with IGP!

intel's chipsets consume craploads of power... 22 watts! ????
AMD A64 x2 4850 +780 G 45 Watt idle. 45 watts cpu
Atom +945 40 watt idle. EHM? 4watts cpu

and still 780G have a 10 times faster IGP minimum i guess, its 6 times faster than GMA3000!

and btw, sorry for being so, dunno the word for it, I'm just releasing my anger from my intel laptop that didnt even play STARCRAFT :mad: starcraft = Direct DRAW 256 colors @ 640x480...

laptop was 2 gb ddr 333 mhz Intel p4 mobile 1.8 ghz Intel 915 chipset.

starcraft lagged, linux went shit slow with cpu usage while using compiz ( IGP suuuucked)
issues with igp all the time.
 
When it comes to a sheer performance comparison, the VIA Nano rules and the Intel Atom drools.
Click to expand...
This is the example of what is the real issue. This is in the conclusion. People will look at this and only take this away. Again, the Power Draw graphs. People will look at that and only take away "Huh, the Atom draws a ton of power for very little work relatively".


The Conclusion should have been more level headed and then you can insert some humour after the the point was made. With the Power Draw graphs, it should have been immediately emphasized that the chipset is skewing the results. What should have been taken away from the Power Draw graphs is that the current platform for the Atom is inferior (and that people looking for an Atom system could benefit from waiting/looking for a better chipset).


The article is construed as Atom vs Nano. Not An Atom System vs A Nano System. This is the problem. It was an extremely informative article, but was written without enough emphasis on telling the reader that the CPUs aren't really comparable. There was no question the Atom would've been monkey-stomped (I can use idioms like this because I'm just a poster) by the Nano. The first paragraph started out well though and the information is there...

...I'm just being a bit too critical but a good site can take its critics right?
 
ChronoReverse said:
...I'm just being a bit too critical but a good site can take its critics right?
Click to expand...

Not a problem bro. Your thoughts are noted.

I think you greatly underestimate our core readership though. I would suggest many are like you that will be reading that article. The article is not put together for the person that buys whatever computer is on sale. It was written for our readers. They are a tad bit smarter than you give them credit for. (Ain't like this is CNet for something! :eek: )
 
Hey Kyle,

Great article - I've been looking at some of the ultraportable notebook solutions for the wife (which must pass the WAF (wife acceptance factor) tests), and being generally useable for web browsing and Office documenting in Windows XP is one of the key things necessary for the WAF to be positive.

We tried out one of the Asus Eee PC's, but it was virtually unusable for the above tasks in Windows XP Home (seemed to be more related to the SSD drive tripping over itself with small file writes), and I'm looking at the MSI Wind notebook that has the Atom processor in it. That being said, did you (or could you) see how the Atom and Nano fare using XP, as the UMPC and ultralight notebook segment will likely continue using this OS for the forseeable future?
 
Schro said:
Hey Kyle,

Great article - I've been looking at some of the ultraportable notebook solutions for the wife (which must pass the WAF (wife acceptance factor) tests), and being generally useable for web browsing and Office documenting in Windows XP is one of the key things necessary for the WAF to be positive.

We tried out one of the Asus Eee PC's, but it was virtually unusable for the above tasks in Windows XP Home (seemed to be more related to the SSD drive tripping over itself with small file writes), and I'm looking at the MSI Wind notebook that has the Atom processor in it. That being said, did you (or could you) see how the Atom and Nano fare using XP, as the UMPC and ultralight notebook segment will likely continue using this OS for the forseeable future?
Click to expand...


Don't get me wrong, the Atom system was "generally usable" under Vista 32-bit. As I noted, maybe I am spoiled by Intel's high end desktop processors. But I am pretty sure it would pass the WAF test.
 
Will there be a Dual-Core Atom or Nano or isn't there any information circulating on this yet?

I would have a pretty hard time buying a Nano if the Atom was Dual-Core with hyperthreading (four threads total) even if the Nano's IPC is better.
 
Will you redo the tests with the newer Intel Poulsbo chipset as the platform, when its released to consumers?
 
What about price? I've seen the Atom based Mini-ITX board for about $75, where the Jetway C7 boards average around $150, unless I'm missing something...
 
ChronoReverse said:
This is the example of what is the real issue. This is in the conclusion. People will look at this and only take this away. Again, the Power Draw graphs. People will look at that and only take away "Huh, the Atom draws a ton of power for very little work relatively".


The Conclusion should have been more level headed and then you can insert some humour after the the point was made. With the Power Draw graphs, it should have been immediately emphasized that the chipset is skewing the results. What should have been taken away from the Power Draw graphs is that the current platform for the Atom is inferior (and that people looking for an Atom system could benefit from waiting/looking for a better chipset).


The article is construed as Atom vs Nano. Not An Atom System vs A Nano System. This is the problem. It was an extremely informative article, but was written without enough emphasis on telling the reader that the CPUs aren't really comparable. There was no question the Atom would've been monkey-stomped (I can use idioms like this because I'm just a poster) by the Nano. The first paragraph started out well though and the information is there...

...I'm just being a bit too critical but a good site can take its critics right?
Click to expand...

Anyone who's serious about their computing to the point that they read processor comparisons will most likely take the time to read the article, or at the very least read the entirety of The Bottom Line, which is typically a great overall summary of the article in a couple paragraphs. Anyone who's not? Well, to be honest, they deserve all the frustration they get for not doing the research. Its amazing how taking a few minutes out of your day to actually do it right the first time saves you so much trouble later on.
 
I'm really glad to see VIA making progress in this area! I've been a fan of their mini-ITX systems since the C3, and they have come a LOOOONG way. Glad to see them so competitive. My Linux firewall right now is a refurb'd Pentium 4, which replaced a C7, but I might be going back to a Nano.

Something not mentioned was price. VIA's board partners consistently gouge us on the price - it's not uncommon for a VIA Mini-ITX board to cost well over $300! :eek: In comparison, the Intel board you used for testing is going for $79.
 
ChronoReverse said:
This is the example of what is the real issue. This is in the conclusion. People will look at this and only take this away. Again, the Power Draw graphs. People will look at that and only take away "Huh, the Atom draws a ton of power for very little work relatively".
Click to expand...

Hmm... Reading the article and:

The fact is that the Nano and the Atom are very different.
Click to expand...

you take the time to read Intel’s pages on the Atom, you will never see it designated to be put into “workhorse” applications like we have done today.
Click to expand...

As we mentioned on page one, the Atom and the Nano are two different animals, but I continually see them lumped into the same category when they simply should not be.
Click to expand...

If you dig a little deeper into Intel’s vision of what Atom is for, you will see that it is very much pushing Atom to be used in the Mobile Internet Devices (MIDs) space.
Click to expand...

Atom is not being widely utilized in what it was truly designed for, and that is UMPCs.
Click to expand...

Conversely, the Intel Atom looks to be best equipped for what it was truly designed for and that is Mobile Internet Devices and Ultra Mobile PCs that are not Windows based. The Atom pulls much less power and is smaller in die size due to its forward-looking 45nm fabrication process and less transistors.
Click to expand...

I think you'd need extremely narrow vision to NOT get the point of the article.
 
I love seeing content creation benchmarks on systems designed for content consumption.

in other words computers designed for playing flash and checking emails, used for encoding and decoding dvd movies.:eek:
 
IS there any chance of seeing a VIA/INTEL Netbook comparison from [H] anytime soon?

Looking at a final decision mid august i think (Wind/EEE PC-80g)
 
Absolutely phenomenal article. While they are very different systems, because of the way they are being packaged (in the netbooks), they will undoubtedly become competitors. I think the last few sentences really get at the point of these systems
The Atom is a great replacement for the old C7, mor MID's (like my Nokia n800), and the Nano is perfect for netbooks. Now, if only OEM's would use them as such instead of throwing an Atom into a full scale computer/netbook.
 
Fantastic comparison.

One thing struck me though. I got to thinking about the Asus EeePC and the Celeron M ULV the originals used. The TDP of those is 5w, and I think are pin-compatible with the Via C7-M. Could you source one of those and throw it into the mix? I would be interested to see if the Atom is much of an upgrade over a 2005 part.
 
Sucks that there is such a huge price difference in the Atom boards and the new Via boards. I want to build a low power WHS but the Atom doesnt have gigabit ethernet and via cost a whole lot more if the prices at logic supply are correct.
 
Overall, I think this was an excellent review, and I agree with most everything that was said.

However,

People don't use mini-notebooks (like the EEE) as their primary systems, they use them to surf the web, check email, type documents, listen to music, and watch videos.

None of these activities really require the power of a Core 2, or even a VIA Nano CPU.

You might note that these activities are in line with Intel's "MID" concept. The reality is, no one wants Intel's MIDs. We already learned this with UMPCs. To have a "full" Internet experience, you need a keyboard (for typing IMs, emails, or forum posts quickly) and a moderate size screen (so that the page fits on the screen).

Now, here's the thing: mini-notebooks like the EEE have really small batteries. Usually 3 or 4 cells. And since they are ultra-mobile (compared to regular notebooks) and cheap (again, compared to regular notebooks), people want to take them everywhere. That means that people want good battery life - at least 4 hours, if not more. Small batteries and long battery life mean that you need ultra low power components.

Atom does that. It is the lowest power x86 CPU with reasonable performance. That makes it perfect for mini-notebooks like the EEE and MSI Wind.

The Nano competes in a different space. It's not ultra-low-power, but it still uses less power than a Core 2 CPU. That means that full-size notebooks with 6-cell batteries can have a balance of performance and battery life. More than 1/4 of the price of a $400 notebook is going to Intel right now, which means that there's a big opportunity for VIA here. Of course, they're also going up against AMD in this sapce.
 
One huge difference between the two offerings I would like to see explored is the discrete graphics option available with the Via motherboard. More specifically:
What video cards could be supported?
Quality of video play back with a GPU handling the work load.
Power usage when using a discrete GPU (obviously this would vary depending on GPU)
The obligatory game benchmarks.
 
CyborgTMT said:
One huge difference between the two offerings I would like to see explored is the discrete graphics option available with the Via motherboard. More specifically:
What video cards could be supported?
Quality of video play back with a GPU handling the work load.
Power usage when using a discrete GPU (obviously this would vary depending on GPU)
The obligatory game benchmarks.
Click to expand...

VIA board is an old EPIA design with a Nano fitted on, not a production product, but that old board has x16 size PCI slot, so you would put pretty much any card in the world on it. The Nano systems we saw running Crysis had 8800 series on them IIRC. The new boards will likely be more forward looking and remember that VIA and NVIDIA are still in bed on this.

ssratt said:
IS there any chance of seeing a VIA/INTEL Netbook comparison from [H] anytime soon?

Looking at a final decision mid august i think (Wind/EEE PC-80g)
Click to expand...

Supposed to have a Wind on the way.

MorfiusX said:
What about price? I've seen the Atom based Mini-ITX board for about $75, where the Jetway C7 boards average around $150, unless I'm missing something...
Click to expand...

Keep in mind that VIA has never had any competition in this market before. Supply / demand is controlled by VIA up till now. That will likely change. Also, the VIA had quite a few more features on it IIRC.

Toytown said:
Will you redo the tests with the newer Intel Poulsbo chipset as the platform, when its released to consumers?
Click to expand...

We will see what happens.

MrMike said:
Will there be a Dual-Core Atom or Nano or isn't there any information circulating on this yet?

I would have a pretty hard time buying a Nano if the Atom was Dual-Core with hyperthreading (four threads total) even if the Nano's IPC is better.
Click to expand...

Yes on both. I think Atom's niche in the mininote will be when it duals out as mentioned in the article.
 
Given that my 2.66GHz Q6600 utilizes a full core to play the 1080p clip, what we saw or rather didn’t see, was not surprising.
Click to expand...

I'd love to see a 2.66GHz Q6600. Where can I get one?

/smartass

Deservedly so. Fixt - Kyle
 
One thing about power draw:

If CPU A uses 10W and finishes a task in 20 seconds, and CPU B uses 5W but finishes the same task in 42 seconds, CPU A will still have used less energy in total for the task.

This has to be taken into account when you look at the power consumption figures. Even if CPU B finishes the task in 35 seconds, it still means it has virtually no advantage over CPU A, while being slower (nodoby likes waiting for their computer to finish some task).
 
I'm not a power user as in terms of high spec'd systems. I stopped upgrading years ago, and stopped at the Athlon XP-M 1800+ (not overclocked) ;P

Power versus permance has always been important.

The mobo that CPU was on kicked the bucket. So I picked up Intel's Atom board, and have been using it for the last couple months. I have to say everything I've thrown at it has exceled beyond my expections.

I dual boot, Xubuntu / Windows XP Pro SP3, linux my main system. But anyways under both environments it works lickity split for what I'm use too. No regrets. Even running XP as a guest OS under Linux with VMware was better than I was expecting.

[H] did a great job with the pre-review. I'm looking forward to playing with the new Nano, if VIA can keep the price down. Intel has shown a great goods for the cost of vs. performance.
 
I like to see articles about this. I think in a few years Intel will have a great product on their hands, but right now Atom is just a stepping stone for something better. It draws to much power for pdas/phones, and its not powerful enough for laptops. I'm very excited about 86x phones in the future. The compatible programs and feature list will be huge!
 
Nice article. Thanks.

Kyle, do have any plan for an article about AMD Puma and Intel Centrino 2 platform? I would really appreciate it if you can write an article on these two platforms compairing the performance in normal office and web usage, HD viewing and light games such as WoW performance, power consumption and battery life, price, availability and etc.
 
I first want to say that I enjoyed the article. I was expecting to see the atom compete with the nano a little more then it did.

In the future if you do more tests like this though if possible throw in the older via chip system into the mix. That way one can see how these compair to the last gen. Also would be nice to see a core-solo or core-duo notebook chip on a mini-itx thrown into the mix just as a comparison.
 
JimmiG said:
One thing about power draw:

If CPU A uses 10W and finishes a task in 20 seconds, and CPU B uses 5W but finishes the same task in 42 seconds, CPU A will still have used less energy in total for the task.

This has to be taken into account when you look at the power consumption figures. Even if CPU B finishes the task in 35 seconds, it still means it has virtually no advantage over CPU A, while being slower (nodoby likes waiting for their computer to finish some task).
Click to expand...

I was thinking somewhat the same.

For instance, on the mp3 test, the Atom would end up drawing more power. By the time the atom finished the Nano would be running idle about 8:30 minutes

So depending on usage the power draw differences may be negligible.
 
When a VIA rep hands you the chips and tells you "Here, do whatever you want", you know they know what's going to happen.

Great article. As for me, I would never buy an Atom-based PC/notebook unless it was only for reading email and browsing the internet. Frankly, I was suprised how well, the Atom did on the encoding benchmarks but dissappointed it could not decode blu-ray.

In case it has not been said enough, the Atom is not meant to replace a Core, a Pentium or even a Celeron (do they still make Celerons?). Intel designed this chip for UMPCs and not ultra-portables: eg. for uber-iPhones and not for eeePCs. If you are looking for an ultra-portable, I would recomend you compare the Nano vs a Celeron vs whatever AMD will roll out.
 
Finally a good article about the true purpose of the Atom. I'm tired of seeing waves of people complaining that the Atom is not in every computer ever made. While the Atom has its uses, I have no use for it.
 
LeftSide said:
I like to see articles about this. I think in a few years Intel will have a great product on their hands, but right now Atom is just a stepping stone for something better. It draws to much power for pdas/phones, and its not powerful enough for laptops. I'm very excited about 86x phones in the future. The compatible programs and feature list will be huge!
Click to expand...

I agree 100%. MIPs is still better on power, while Nano has more horsepower. I would love to see how a 45 nanometer Nano would compare to the Atom. Atom is a pretty good first attempt, but it's still got a long way to go... the future will be interesting!
 
In the end There's one more thing that needs to be considerd given an specific. Before I post that thing we must take into account that a common user drains power for at least 8 hrs a day. With that said and considering the real time video encoding tests, where Kyle compares the Nano to the Core 2 Xtreme 9750 one must asks. What is the total kWh comsuption difference between the Core 2 Extreme and the Atom/Nano, for the 800 mins tests. If the Core 2 takes about 60 minutes and consumes 4 times more power I think that given X as the power consumption by minute 60min*4X=240X < 800X so then the Core 2 Extreme is more efficient. Of course this judgement is given in the frame of the tests shown. Maybe with time some "specific" tests for ultra low power mini CPUS would be a better approach.
 
You must log in or register to reply here.
Back
Top