The benchmarking procedure (the list of software and test conditions) is described here. To make the diagrams easier to read, results are represented in percents (100% stands for the result of Intel Core 2 Quad Q9300 in each test). Detailed results in absolute values are provided in this spreadsheet.
Multithreading is still weak in interactive operations in 3D Modeling programs, and it's implemented only in some of them. So it would be logical to assume that dual-core processors will take the lead here, probably accompanied by the fastest triple-core processor. That's exactly what happens, real differences between the first three processors are not very big. If we take a look at detailed results, we'll see that Athlon II X2 250 leads in SolidWorks, Athlon II X3 435 -- in Lightwave, and Pentium E6300 leads in the other programs.
What concerns rendering in the same applications, both dual-core processors are outperformed radically even by the triple-core processor from the old Phenom line. And in their turn, the new triple-core processors apparently occupy the golden mean niche here. Especially the 435, which performed well in the interactive test, was a little outperformed by the quad-core processor in rendering, but managed to catch up with the processor from Series 700 (but it was handicapped by DDR2 memory). On the other hand, the 435 is apparently cheaper than the 710. And if it continues to demonstrate the same good performance in the other tests, its fate will repeat that of the Phenom II X4 810, which had been pushed out by Athlon II X4 620/630.
This is another group of applications, which are not optimized for multithreading good enough. As a result, the first place is taken by the dual-core processor with the highest frequency, and the second place is shared by a couple of triple-core processors (the 435 did well owing to its relatively high frequency, while the 710 succeeded thanks to its large cache, welcome by such tasks).
Compilation speed is affected by each additional core, so the overall principle is crystal clear: the more cores, the better results. However, memory has a noticeable effect as well: the large cache helped the 710 to get very close to the quad-core processor, and transition from DDR2 to DDR3 has the same effect on all new triple-core processors as the 200MHz difference in clock rates.
We've traditionally registered the advantage of Intel processors in this group of tests. So our comparison of Athlon II X2 250 and Pentium E6300 (equal competitors in general) showed AMD's defeat in this group. However, the third core in combination with relatively high clock rates of the new processors opposing the E6300 make up for this drawback. The average score, compiled from such promiscuous applications, does not give an adequate picture on each one separately. So let's take a look at the detailed results.
|Runtime (lower is better), mm:ss
||Phenom X3 8750 (DDR2)
||Athlon II X2 250 (DDR3)
||Athlon II X3 425 (DDR2)
||Athlon II X3 425 (DDR3)
||Athlon II X3 435 (DDR2)
||Athlon II X3 435 (DDR3)
||Athlon II X4 620 (DDR3)
||Phenom II X3 710 (DDR2)
||Pentium E6300 (DDR3)
Pentium E6300 still leads in ACDSee, PaintShop Pro, and PhotoImpact. No wonder, these three applications cannot take advantage even of two cores. They can use only one core in most operations. And in this situation it's good to have the entire L2 cache allocated to a single core.
On the other hand, even the slowest of the new triple-core processors with DDR2 defeats its competitors in graphics editors Paint.NET and Photoshop, which can rationally distribute the load between cores.
Just like in the compile test, we can see the strict arrangement of processors according to the number of cores. Processors with more cores take up higher positions on the diagram. It also concerns the old Phenom X3 8750. Clock rate stands on the second place in its effect on test results. So the new triple-core processors lined up evenly above Phenom X3 710, which large cache helped it not to lag behind Athlon II X3 425, which clock rate is higher only by 100 MHz.
Write a comment below. No registration needed!