|
It is not huge as in breakthrough what AMD has accomplished with the C3 Revision but it is still worth writing home about it. The two things that are noteworthy are the increased overclocking potential at "slightly altered stock settings" and the pretty dramatically reduced power consumption of the C3 revision. It is difficult to judge how much of this is caused by the hardware-based power management compared to the earlier software/firmware-base power management but the results definitely speak for themselves.
In the same context, it is absolutely necessary to stress the point that benchmarking scenarios are probably the worst case to hi-light the accomplishments of the new power management because we are running either at idle or else at full load. Consider just the situation of typing this article, where the input is frequent enough to keep the CPU from entering the lowest power states, simply because the older scheme was too slow. This is where the new, CPU hardware-based power management can make a much bigger dent in the overall power budget than what we can show in our benchmarks, since the CPU will enter idle states much more frequently than before, and do so without causing the infamous "CnQ performance bug" to surface.
In so far, we are probably under-appreciating the overall achievement of the C3 stepping. At the same time, there is, however, still a bitter pill.In the mobile sector, AMD has very successfully implemented a throttling of the NB under idle conditions. We sort of anticipated the same thing to happen in the desktop space with the C3 stepping, alas, nothing on this front seems to have happened. Knowing how much power the NB consumes, this could be another significant power savings feature if brought to the desktop.
|
RSS Newsletter