Intel is adding a new processor to its line of Cedar Trail chips for low power desktop computers. The 1.86 Ghz Intel Atom D2550 will be a lot like the Atom D2500, but with a few important tweaks.
First, the D2500 doesn’t support hyperthreading, while the D2550 does. That means that some software task managers will show the new chip as having 4 cores, even though it’s actually a dual core chip. The upshot is that apps that can take advantage of hyperthreading should run more quickly and efficiently.
I was actually a little surprised when cybergusa pointed out to me that the D2500 doesn’t feature hyperthreading capabilities, since most Atom chips since the original Atom N270 have had that feature. But I checked with Intel, and it turns out it’s true.
The second change is that the Atom D2550 processor will have a graphics clock speed of 640 MHz compared with 400 MHz for the Atom D2500 processor.
The rest of the specs remain the same, with both processors featuring 10W TDPs and 512KB of L2 cache.
According to VR-Zone, the new Atom D2550 processor could launch on February 26th and begin shipping soon after.
Intel also offers a faster 2.13 GHz Atom D2700 dual core processor with 1MB of L2 cache, 640 MHz graphics, hyperthreading support, and a max TDP of 10 watts.
None of these processors are designed for high-performance systems, but they should provide a little more bang for the buck than the latest Atom N2600 and N2800 processors for netbooks.
Can someone explain what makes this a “desktop” CPU?
Chips are either optimized for mobile, maximizing power efficiency for better battery run times, or for max performance. The later is what’s meant by desktop or in this case nettop CPU.
Features like Intel SpeedStep, which auto reduces clock speed when system is idle to save power, are examples of what they put into mobile chips but not the desktop ones.
This also means the desktop/nettop chips are generally clocked higher and are free to operate at higher wattage. Though even these chips are low powered compared to Intel’s higher end offerings like the Core i-Series. It’s just for the ATOM line they are intended for less mobile usage.
So to summarize, for Intel ATOM’s the N-Series is their mobile line and the D-Series is their desktop/nettop line. They also have the Z-Series, which goes towards their ultra mobile and embedded system offerings like Oak Trail (Tablets and embedded systems) and Medfield (Smart Phones and some Tablets).
Btw, for AMD Fusion there is a similar differences with their Ontario C-Series competing with Intel N-Series, the Zacate E-Series competing with Intel D-Series, and the Desna Z-Series that’s optimized for Tablets.
Because of the higher power and thermals of the Desktop/nettop range chips. You won’t find them put into anything smaller than a 11.6″ screen system.
Thanks! Is Intel SpeedStep like low-TDP or something different?
How much more energy would the D-series use at idle than the N-series?
Intel Speed Step is like throttling a engine. Lower speed when idle
allows for lower power usage. While allowing it to return to full speed
when performance is needed.
Without it CPU’s stay at default state and doesn’t throttle.
Unfortunately, they don’t usually publish the detailed power consumption
range of each chip. So can’t tell you how much more energy the
D-series would consume versus the N-series while idle, but even a small
difference can add up over time.
Hey, Brad, there may be a minor typo: The D2500 that has a 400mhz GPU; the D2550 has (presumably) a 640mhz GPU.
Here’s the ark.intel link on the D2500:
They don’t have an ark link for the D2550 yet.
I also wonder if any of these secretly have VT-x the way the N570 had VT-x, even though ark.intel claims it didn’t. I assume Intel added VT-x to the N570 to make Seamicro happy, but kept quiet about it so that other people looking for a low-cost server would use a more expensive CPU that officially has VT-x.
No typo, that’s what Brad indicated for the GPU clock speeds.
While Intel usually doesn’t publish product pages until it actually officially released. So much of the information is drawn from what Intel has indicated and any data sheets with extra information.
Intel also tends to cripple features instead of actually omitting them from the chip design. So VT-x was possible on the N570 as long as you had a BIOS that didn’t disable it.
Though whether it works 100% is another thing. ATOM architecture is still far behind the rest of Intel’s offerings and won’t receive the major update until the 22nm Silvermont.
Like for example Asus managed to get memory remapping working on the 1215N with a BIOS fix but that doesn’t mean it has full 64bit memory handling performance.
While it’s presently unknown if the same BIOS trick will work with any of the newer Cedar Trail chips.
Comments are closed.