Non-professional tests
Archiving
Quite vivacious, but apparently insufficient. Results of a single-threaded UltimateZip (see the Excel file) suggest sad ideas that the problem is not in complex communications between cores, but in the computing power of the core - it's lower than that of the Intel Core architecture.
Encoding
This couldn't be any worse. Phenom is outperformed even by the dual-core Core 2 Duo E7200, although this group includes some tests that can use four cores.
Games
Games do not bring good news either.
Non-professional photo processing
That's a shame. As before, Phenom is outperformed in absolutely all tests in this group. The champion is faster than this processor by 12-32% (that is almost by one third!)
Total non-professional score
Despite the cardinal change of software genres, the average score is surprisingly stable: it's again similar to that of Core 2 Duo E6600. Even the E7200 is doing better...
Estimated power consumption*
* We actually measure power consumption of the on-board VRM, so our readings may be higher, because VRM does not have the efficiency factor of 100%.
Idle
100% load
Unfortunately, the situation is far from peachy here. We can try to explain high power consumption under 100% load with the old 65-nm fabrication process and the lack of experience in quad-core designs (Phenom is AMD's first attempt), but the remarkably high power consumption in the idle mode is unexplainable. But it's a fact. However, old AMD Athlon 64 X2 processors had nothing to boast of compared to Intel Core CPUs either...
Conclusions
Just 91 points. AMD Phenom X4 9850 performs on a par with Intel Core 2 Duo E6600. These words sound like a hammer nailing down a coffin lid: the top quad-core processor from AMD, launched in 2008, demonstrates average performance on the level with a dual-core processor from Intel (!), launched about two years ago, which occupies the 14th place (!) in the official Intel rating. AMD will play its usual games with lower prices, of course - it has no other choice with this performance level. It's another question that I don't know the answer to: what's next? The K10 core is apparently weak. Besides, it was launched several years after Conroe. That is Intel enjoys a double advantage - in performance of the existing solution and in time to design a new one. I don't know how AMD will find a way out of this situation. But contemplating our test results, a bitter joke comes to mind - the main unfixable mistake in Phenom is this chip itself.
