Tests
For your convenience the results are represented in percent, 100% being the result of AMD Athlon II X4 620. Absolute results are provided in this Excel spreadsheet.
3D visualization
As you can see, the junior dual-core Core i5 doesn't lag much behind the higher-end quad-core processors. Which means the number of cores isn't that important in such tasks. A formal comparison of processors producing close results indicates that clock rate is the most (or the second most) critical parameter in this test. Just like the good old times! But then why do hexacore solutions perform slower than quad-core ones? Turbo CORE should have worked and boosted the clock rate to the maximum, so Phenom II X6 1075T would perform similarly to Phenom II X4 970. It seems this benchmark has more active processes than a regular single-threaded task, and transferring those between free cores creates additional burden.
3D rendering
In rendering tasks the situation changes to the opposite. The processor with most cores wins. Even the junior hexacore CPU with a relatively low clock rate outperforms the competing quad-core solution.
Scientific calculations
Applications in this group of tests vary in terms of load type, so if you're interested in anything particular, download the complete spreadsheet (see above). However, the average result is qualitatively similar to that of the first benchmark. Just the Intel CPUs perform a bit better, especially the Core i5-760.
Image editing
Intel processors lead in this test for known reasons. These applications are both poorly optimized for multithreading (except Photoshop) and mostly depend on the performance of integer blocks. In such situations AMD processor need a large clock rate advantage which they don't have. As for the prospects, not much will probably change until Bulldozer-based solutions arrive. Note the interesting results of Phenom II X6 1075T that catches up with Phenom II X6 1090T. It's probably the effect of certain optimization introduced into BIOS since the release of the hexacore series. Because we have already tested Phenom II X6 1090T before. On the other hand, Phenom II X6 1055T performed rather poorly.
Data compression
Different archivers prefer different processors: 7-Zip likes AMD solutions, while RAR favors Intel's. Since RAR produces two results in this benchmark (packing and unpacking), the bottom line shifts greatly to the Intel's side.
Compilation
This benchmark gives hexacore processors a chance to win back. And more than in rendering tasks at that, especially if you compare the results of the junior solution. This benchmark is both multi-threaded and quite critical to memory subsystem efficiency. As you can see, 8MB L3 hardly help Core i5-760 make up for the smaller L2 comparing to Phenom II X4. At the same time, the 4MB of Core i5-655K clearly helped it outperform Athlon II X4 models without L3, although just by a neck.
Write a comment below. No registration needed!