SuperPi, Pifast and WprimeThe most popular 2D benchmarks are of course these three, SuperPi and Pifast because of the simplicity of the benchmark and Wprime because it's able to spread the calculation progress over multiple cores. We've added all three benchmarks in two graphs: one for the fast calculations and one for the more stability related calculations.
As you can see, the extra L2 cache provides an extra performance boost, an overclocked E6300 at 3500MHz is just slightly faster than a stock clocked E8500 and loses terrain when the E8500 is set at the same speed. When overclocking even more, we can see that the E8500 uses the extra MHz’s well enough to scale further and produce far better scores.
Again, the L2 cache pays of as we see the same differences occur as we have seen in the previous graph.
Sisoft SandraSisoft sandra cpu bandwidth benchmarks such as Dhrystone and Whetstone give us an indication of the theoretical performance and how these compare to other processors.
As you can see, the 45nm series outperform the 65nm at same frequency and even at higher frequencies. A stock E8500, clocked at 3,16G, manages to stay ahead of an overclocked E6300, remarkably.
WinrarTime to find how the 45nm perform in real-life applications, such as Winrar, which is widely uses to compress files. The higher the compression rate, the faster you file will be compressed.
We see things we'd never expected to see: at stock frequencies, an E8500 is not at all faster than its 65nm competitor at overclocked settings; however at same frequencies the situation is the opposite.
On a side note, to give you an idea how to interpret these numbers, I'll give an example of how fast a 1GB folder could be compressed when the E8400 is overclocked to 470x9,5=4465MHz.
- 1GB = 1024MB = 1024x1024KB = 1048576KB
- Compression rate: 1933KB/s => 1048576KB in approx. 542 seconds
- 542 seconds =
9min 2s