Phenomenon Redux: AMD Phenom II in Linux Test
A recent Linux Community test compared the AMD Phenom X4 with a Turion X2 and Intel P9400. Meanwhile AMD has sent us its new Phenom II processor, prompting a new test. These are the latest benchmark results.
The Phenom II Z940 processor is officially clocked at 3 GHz, which the Linux Kernel 2.6.27 recognized, although /proc/cpuinfo reported the model itself as still unknown. The new Phenom can be lowered to as much as 800 MHz via speed-stepping. When idle, the CPU in our test system used up 65 watts of power, while its 2.2-GHz Phenom X4 9550 predecessor required about 75 watts.
The Phenom II footprint is indistinguishable from that of the Phenom I, so that it can also fit into the mainboard's AM2+ socket. The Phenom 9550's internal cache is 2 MBytes, while the Z940's is three times that at 6 MBytes. The Linux tests involved an Asus M3A78-EM mainboard with a 1-GByte main memory.
Our test used the same ffmpeg benchmark as before, whereby a script had to convert one, two, three, four, six and eight AVI files into H.264 format. Apart from speed, we also measured power consumption. The following graphs show the results.
The new quadcore CPU proved to be around 30% faster than its predecessor, but also clocked much higher. It went nose-to-nose with the Intel P9400 in processing speed until about two concurrent tasks, then sprinted ahead. Power usage, on the other hand, was somewhat less impressive.
The Phenom II might use somewhat less idling power than the 9550, but still considerably more than the Intel Core2. At full load the Z940 test system required almost 160 watts.
The new Phenom is clearly better than its predecessor. It uses less power when idling and still provides better performance. However, its higher purchase price brings it closer in line with the Intel CPU. Another advantage is its backward compatibility so that the CPU can fit on any current AM2+ boards. Even a couple of older AM2 boards can accommodate it after a BIOS update. As far as the competition, Intel is promising its own i7 test system in early February.
|Gallery (4 images)|
DetailsYes, we measured the power of the whole system at the outlet and the CPU was running at 800 MHz (dynamic frequency scaling), when consuming only 65W.
Thanks for this article!
Could you please post details of your test system?
Was the power usage at idle (65W) meansured at AC outlet? Was the processor downclocked by cpufreq or it operated at full speed?
Thanks in advance!
Kernel king admits his tone has alienated volunteers, but says the demands of the process require directness.
New flaw in an old encryption scheme leaves the experts scrambling to disable SSL 3
Lennart Poettering wants to change the way Linux developers talk to each other.
Enterprise giant frees itself from ink and home PCs (and visa versa).
Mozilla’s product think tank sinks silently into history.
TODO group will focus on open source tools in large-scale environments.
New tool will look like GParted but support a wider range of storage technologies.
New public key pinning feature will help prevent man-in-the-middle attacks.
Carnegie Mellon researchers say 3 million pages could fall down the phishing hole in the next year.
The US government rolls new best-practice rules for protecting SSH.