||AMD Phenom II X4 965
||AMD Phenom II X6 1090T
||Intel Core i7-930
||Intel Xeon X5680
|Core clock (std/max), GHz
|Cores (HT threads)
|L1 cache, I/D, KB
|L2 cache, KB
||4 x 512
||6 x 512
||4 x 256
||6 x 256
|L3 cache, KB
||Gigabyte 890FXA-UD7 (AMD 890FX)
||Corsair CM3X2G1600C9DHX (2x DDR3-1333, 7-7-7-20-1T, unganged mode)
||Intel DX58SO (X58)
||Kingston KVR1333D3N9K3/6G (3x DDR3-1333; 9-9-9-24)
- Seagate 7200.11 (SATA 3Gb/s) HDD;
- Zalman CNPS9700 cooler;
- PowerColor HD5870 1GB GDDR5 graphics card;
- SeaSonic M12D 750W PSU.
Since our tests coincided with the update of our test method, we chose the necessary and sufficient minimum of competing CPUs. Of course, it was interesting to find out how AMD's new top-end CPU would compare to the previous leader AMD Phenom II X4 965. As for Intel solutions, we couldn't go without a similarly priced rival. Formally, there are two rivals -- Intel Core i7-860 and Intel Core i7-930, -- but the latter is positioned closer to the novelties. Besides, we were to test AMD CPUs with a motherboard based on the top-class chipset. So it was only fair to compare it to Intel's top-class LGA1366 platform as well. Finally, we needed a recent hexacore CPU from Intel, but we didn't have any Intel Core i7-980X lying around, so we took Intel Xeon X5680 of nearly identical characteristics. Though, in terms of price, neither of Intel hexacore CPUs can compete with any of AMD novelties. You can build a whole Phenom-based machine for the price of a single Intel Core i7-980X CPU.
As usual, the absolute test results are provided in the summary, while the diagrams below are relative to the reference CPU considered 100%. The current reference CPU is Athlon II X2 630, so you can see how much the current mid-end solutions are outperformed by the novelties.
Note that the final version of the test method (to be fully featured in the next processor-related review) might have more tests than this variant. Today, we used a kind of beta version. The word "beta" was quite adequate to the tests results we got, by the way. Read on to find out why.
This group of tests was the only one, where AMD Phenom II X6 1090T yielded to AMD Phenom II X4 965. As we all know, applications of this kind are still badly optimized. Besides, newer versions of those didn't change much. But why the Turbo mode didn't help? Well, there may be two reasons. Firstly, the active use of one or two cores is not the same as not using other cores at all. So, perhaps, some non-resource-critical processes were periodically assigned to free cores, preventing those from entering the passive mode. These 3D suites are quite complex after all. Secondly, and most likely, that was just the BIOS fault. Especially as we had to use a pre-release version. Remember the first tests of Intel Core i7, especially those of Core i7-920? Initially, it was given the cold shoulder for performing like a mid-end Core 2 Duo CPU. But when BIOS was fixed, everything fell into place. No wonder, because technologies like Turbo Boost strongly depend on management logic.
Anyway, if we take a look at the detailed results, you can see that AMD Phenom II X6 1090T could only lead in UGS NX. But this test had always been very close to single-threading. Was it because Turbo CORE did its job right? The results of Intel CPUs in UGS NX were suspiciously low, though we rechecked them several times. But even the visual comparison indicated that the Phenom-based testbed worked faster and finished earlier. That was strange, we had never seen this benchmark to behave like that.
The two additional cores were used in all three rendering tasks. For this reason, AMD Phenom II X6 1090T confidently outperformed its quadcore counterpart. However, that wasn't the largest performance difference, you should see more test results below. As for comparison with Intel CPUs, AMD Phenom II X6 1090T won in Lightwave and Maya, while the quadcore (octa-threaded) Intel Core i7-930 won in 3dsmax. All in all, it was a parity.
In this group of tests AMD Phenom II X6 1090T was sometimes helped out by the additional cores, sometimes by Turbo CORE. Anyway, the new CPU didn't outperform AMD Phenom II X4 965 much (perhaps only in Mathematica). As for Intel CPUs, it was a parity as well.
Adobe Photoshop was the only application with full-fledged multicore support in this group of benchmarks. It tried to use six cores, but couldn't do it in every task. For this reason AMD Phenom II X6 1090T is slightly behind Intel Core i7-930. However, in ACDSee the novelty experienced a surprising performance boost. This was hardly due to efficient multithreading, most likely it was the effect of Turbo CORE. The remaining two image editors never cared about multithreading much, so AMD Phenom II X6 1090T performed on a par with AMD Phenom II X4 965. This determined its place on the diagram.
Archivers didn't even try to use six cores, so AMD Phenom II X6 1090T could only hope on Turbo CORE. Well, it only helped out in 7-Zip. This obviously indicated some leeway for further, more aggressive optimizations of this technology.
Thanks to the decent multicore support of the Microsoft's compiler, AMD Phenom II X6 1090T confidently outperformed two closest rivals.
Write a comment below. No registration needed!