Anyone know anything about speed benchmarks for the pineview atom?

[boed]

I know about the graphics and the power consumption. I just set up a Vaio X laptop for a client and it was very sleek but too slow at 2GHz. Any news on faster atom procesors or 64bit Atom processors? I heard the new pine view atoms are only going to be 1.6GHz and 1.8GHz.

 

[hans007]

from what i read the new atom just has all of the stuff on the same die. the core might not be changed much from the current one other than it is 32nm with the igp and northbridge combined.

also there are 64-bit atom cpus and they are no faster, it wouldn't help you. all the atom chips in desktops are 64-bit and it makes no real difference.

 

[21stHermit]

Pine Trail Atom is still 45nm, no or little speed increase. Even Cedar Traill Atom on 32nm a year from now will still be 1.6GHz. Intel is going lower power with each node, not faster.

I'm pretty sure my Atom 330 is 64-bit but I'm running an old XP. I've been very pleased with Atom 330. If you need more go Clarkdale.

 

[boed]

I appreciate how small and low power it is. The only thing that I can say is that I'd trade off a little power consumption for a little more speed. The Sony Vaio X is fantastic for many things but even booting on ssd with a 2GHz atom it is wretchedly slow. I understand I'm asking for my cake and want to eat it too but that is what I want. I'd rather they keep the power consumption as much as the current atoms but with 50% more speed than 50% lower power consumption.

 

[n7]

I hate atoms for real world usage on notebooks/netbooks.

ULV C2D is the way to go IMO.

 

[IntelUser2000]

I appreciate how small and low power it is. The only thing that I can say is that I'd trade off a little power consumption for a little more speed. The Sony Vaio X is fantastic for many things but even booting on ssd with a 2GHz atom it is wretchedly slow. I understand I'm asking for my cake and want to eat it too but that is what I want. I'd rather they keep the power consumption as much as the current atoms but with 50% more speed than 50% lower power consumption.

Pineview doesn't have much lower TDP. It's only in Moorestown the TDP will go down. It goes from Diamondville's 2.5W CPU/4W MCH/3W ICH to 6.5W CPU that includes the MCH and a 2W ICH. Of course, the Menlow platform in Vaio X is already lower TDP than Pineview.

(Mind you 6.5W CPU in Pineview is 2.5W CPU + 4W MCH)

The integrated memory controller on Pineview supports DDR2-800, which is 50% increase from DDR2-533 on the current Diamondville Atoms. The lower latency to memory WILL help the 2-wide, in-order chip quite a bit, but with lower initial clock speed the gain will be only 10-15%.

 

[21stHermit]

Pineview doesn't have much lower TDP. It's only in Moorestown the TDP will go down. It goes from Diamondville's 2.5W CPU/4W MCH/3W ICH to 6.5W CPU that includes the MCH and a 2W ICH. Of course, the Menlow platform in Vaio X is already lower TDP than Pineview.

(Mind you 6.5W CPU in Pineview is 2.5W CPU + 4W MCH)

The integrated memory controller on Pineview supports DDR2-800, which is 50% increase from DDR2-533 on the current Diamondville Atoms. The lower latency to memory WILL help the 2-wide, in-order chip quite a bit, but with lower initial clock speed the gain will be only 10-15%.

You must design systems to know all this.

 

[cbn]

Why are these tiny mobile cpus transitioning to 32nm slower than Clarkdale/Arrandale?

Wouldn't it be easier to make such a small cpu on 32nm?

Or is this a situation where demand and the ability to charge a price premium is a greater factor in the decision making?

 

[21stHermit]

Why are these tiny mobile cpus transitioning to 32nm slower than Clarkdale/Arrandale?

Wouldn't it be easier to make such a small cpu on 32nm?

Or is this a situation where demand and the ability to charge a price premium is a greater factor in the decision making?

IMO, Intel wants to get into new markets, like smartphones. Currently all Intel x86 CPUs take way to much power. So the only way to get there is to keep driving power down with each node, not increase performance.

Intel certainly doesn't need to compete with itself and make a Clarkdale capable Atom. I've been using Atom 330 as my primary PC since May, only the graphics leaves me wanting. CorelDraw almost won't run.

Keep in mind the Atom die is significantly smaller than any other x86 CPU. So even at its low price it is VERY profitable for Intel. Lots of em on a 300mm wafer.

 

[IntelUser2000]

Why are these tiny mobile cpus transitioning to 32nm slower than Clarkdale/Arrandale?

Wouldn't it be easier to make such a small cpu on 32nm?

Or is this a situation where demand and the ability to charge a price premium is a greater factor in the decision making?

Well, the 45nm is a proven, mature process while 32nm is not. It's probably more cost-effective for them to allow their mainstream CPUs to be on the cutting-edge and when the process tech becomes mainstream, transfer that to the Atom.

It costs billions of dollars to transition to the new tech, Atom is cheap to make, but might not help their bottom line too much. After all, the biggest goal here is to make money.

It's more important for Atom to reach lower power, which their 45nm SoC process with much lower leakage will help more than their 32nm high-performance process. That also factors a lot into 1 year late transition since their SoC process is a year behind their HP process.

 

[boed]

I hate atoms for real world usage on notebooks/netbooks.

ULV C2D is the way to go IMO.

Yeah, while even I had to admire how light and slime the vaio x was, it was too easy to max out the processor. Even copying files over the network (lots of them) maxed out the CPU.