When will Intel develop and release a decent low power budget end CPU? The Atom is not much better than a Tualatin Pentium 3. Clock for clock it isnt better, actually worse. And P3 Tualatin is 2001. Come on Intel! Pull your finger out!

Once Intel has a much better budget CPU, then all these nettop devices will be exciting. Until then, they are lackluster.

OEMs should ditch Atom. Just build a nettop using laptop components.

Completely BonkersWhen will Intel develop and release a decent low power budget end CPU? The Atom is not much better than a Tualatin Pentium 3. Clock for clock it isnt better, actually worse. And P3 Tualatin is 2001. Come on Intel! Pull your finger out!

Once Intel has a much better budget CPU, then all these nettop devices will be exciting. Until then, they are lackluster.

OEMs should ditch Atom. Just build a nettop using laptop components.

These CPU's are FAR from P3 status.

www.cpubenchmark.net/cpu_lookup.php?cpu=Intel+Atom+N2800+%40+1.86GHz

Completely BonkersWhen will Intel develop and release a decent low power budget end CPU? The Atom is not much better than a Tualatin Pentium 3. Clock for clock it isnt better, actually worse. And P3 Tualatin is 2001. Come on Intel! Pull your finger out!

Once Intel has a much better budget CPU, then all these nettop devices will be exciting. Until then, they are lackluster.

OEMs should ditch Atom. Just build a nettop using laptop components.

For the first Atoms this was true, but not anymore, as brandon said. They're still slow, but that is kinda the idea.

Nice. Show a multithreaded benchmark, then compare the multi-core multithreaded CPU vs. the single threaded one.

OK, lets take the current gen Atom (but single core this time) and compare it to P3.

www.cpubenchmark.net/cpu_lookup.php?cpu=Intel+Atom+D2500+%40+1.86GHz
www.cpubenchmark.net/cpu_lookup.php?cpu=Intel+Pentium+III+family+1400MHz

P3 is better clock for clock. QED.

Comparing the D2800 to the P3 would require you to FAIRLY compare it to a server III-S twin CPU. So double the P3 scores = 640+ to the 700+ scores for the Atom and the P3 still wins clock for clock.

but that is kinda the idea

No! Please. Slow is not the idea. Budget and low power is the idea. AMD does a better one BTW, or would you say AMD has got it wrong by making it too fast? :banghead:

REPEAT
Atom is long in the tooth already. intel needs to a new faster gen of Atom out. A2xxx is a very poor show given that Atom is already 4 years old, and based off P3 that is 12 years old. Our friend Moore and his Law would say the Atom should be a factor of at least (between) 12/2=6 and 12/3=4 or 2^4=8x or 2^6=64x as fast as P3 or some better combination of power/performance.

You're looking at this from the wrong perspective. A low-power component's first priority is just that: low power consumption. It isn't measured in terms of raw performance; it is measured in performance per watt. This is where the Atom D2500 trounces the P3 that you're comparing it to. :nutkick:

Oh, and Moore's "law" is actually a rule of thumb (definition: "a principle with broad application that is not intended to be strictly accurate or reliable for every situation").

^^ what are you talking about?! P3 = 30W. Atom 10W. An improvement, but hardly trouncing P3 given 12 years.

Moore's "law" is actually a rule of thumb. Congratulations! You demonstrate you can understand "the law". So consider that it might mean during the last 12 years... and what it has meant in terms of other CPUs even CPUs made by Intel. Then apply this knowledge to the Atom!

If you think Atom is a marvellous feat of engineering, that is pushing the envelope of technology, then step right up and say that. I have a different view, that it is significantly behind what our expectations are.

Completely Bonkers^^ what are you talking about?! P3 = 30W. Atom 10W. An improvement, but hardly trouncing P3 given 12 years.

The P3 draws 3 times as much power, but is it 3 times faster than the D2500? No. It's performance per watt is much worse. Now do you get the picture?

Completely BonkersMoore's "law" is actually a rule of thumb. Congratulations! You demonstrate you can understand "the law". So consider that it might mean during the last 12 years... and what it has meant in terms of other CPUs even CPUs made by Intel. Then apply this knowledge to the Atom!

Moore's law does not take power consumption into consideration. It is only concerned with raw processing performance. Using your "logic", I should complain that the Ford Mustang gets such lousy gas mileage compared to the Toyota Prius. The Mustang was designed for performance, and the Prius was designed for fuel efficiency. Two different goals, two different end products.

Completely BonkersIf you think Atom is a marvellous feat of engineering, that is pushing the envelope of technology, then step right up and say that. I have a different view, that it is significantly behind what our expectations are.

I never claimed that the Atom was a speed demon or a wonderful design. But in that power range, it's a decent CPU. You keep ignoring that little detail.

I refer you once again to what I posted earlier. Combine this with what you just said: Atom is 3x more power efficient than P3-S, but computationally not faster per clock. So I ask you the question, is 3x an appropriate improvement for a 12 year period?!

At no time did I say that P3 was better than Atom. What I said is that in the prevailing 12 years we should see (with judicious use of Moore's approximations) about a 2^6 improvement or 64x better. That can be a combination of clock speed, power efficiency, and raw computational performance. But we don't see the Atom as 64x better than a P3-S. And hence my statement, our expectations are higher and Atom is NOT pushing the envelope. Room for improvement Intel.

I think we should stop this fruitless discussion - we are derailling "J&W Intros M001 Nettop" thread.

Using Moore's law in discussions like this is like Godwin's law. Moore only observed the amount of transistors on a die that could be inexpensively produced. Has nothing to with performance. It also doesn't mean every single chip produced needs to have the maximum amount of transistors.

Completely BonkersWhen will Intel develop and release a decent low power budget end CPU? The Atom is not much better than a Tualatin Pentium 3. Clock for clock it isnt better, actually worse. And P3 Tualatin is 2001. Come on Intel! Pull your finger out!

Once Intel has a much better budget CPU, then all these nettop devices will be exciting. Until then, they are lackluster.

G530, $42 compared to $42 of the D2500 (in 1000s). Question you should be asking is "Why has no one introduced a cheap SB based fanless nettop?" The answer is that TDP of G530 is 6.5x 2500, therefore its almost impossible to make it fanless.

620T is 35w $70 if you are interested