Intel Dynamic Acceleration
The idea behind Intel’s “legalized” dynamic overclocking tool is very simple. When one of the cores of a dual-core processor switches to power-saving C3-C6 state, the clock speed of the second core increases by 200MHz. At the same time, the processor power consumption and heat dissipation do not get beyond the set limitations, because one of the cores remains in idle mode. This technology is absolutely transparent, it works on the CPU and chipset levels and doesn’t require any additional drivers in the system.
It looks something like that:
When a single-threaded application loads fully one of the processor cores of the dual-core CPU, the processor clock frequency multiplier gets 1x higher than its nominal value.
It sounds very attractive. Looks like Intel has finally come up with a way of speeding up single-threaded applications processing on dual-core CPUs. But unfortunately, all these statements are missing one important point. For Intel Dynamic Acceleration to work effectively, the second core needs to be in Deep Sleep (C3) or even deeper mode. It means that the second core shouldn’t be running any background processes at this time. Unfortunately, practice shows that this is a very rare occasion, at least in Windows Vista. When one of the cores gets loaded with work, the OS transfers all background processes to the second core preventing it from going into Deep Sleep mode. For example, we had to manually transfer all processes to the same core in order to be able take the screenshot you have just seen above.
Therefore, you shouldn’t really hope for Intel Dynamic Acceleration technology to ensure a significant performance improvement during single-threaded applications processing.