Where does the process begin?
In the case of the processor, the most common technique consists of making it work at a faster speed than it says. Microprocessors are designed in accordance with a range of working speeds that is more or less established from the start. This design is made according to electronic (paths of the electrical current through the chip) and thermal criteria (the maximum heat the chip is able to dissipate).
|Unfortunately, the engineering and manufacture of chips are not exact sciences and it is impossible to know in advance the exact speed at which any particular chip will be able to work. Once a series of processors has been obtained (in principle, all at the same established speed, e.g. 800 Mhz), these chips are tested in a test bed. Those passing the tests at 800 Mhz are labelled as such and put on sale. || |
Those that don’t pass the test are successively transferred to tests at lower speeds (e.g. 766 or 733 Mhz) and are labelled with the speed of the test they eventually pass. These tests are extremely demanding and, to carry on with the example, we may find a processor on the market originally designed to work at 800 Mhz but labelled at 733, which gives us “ample” room for overclocking.
The level of overclocking of a processor will depend on the specific make and model; some really cannot be pushed any further, or may only be stretched a small amount. Generally speaking, component speed can be increased by an average of 15%, with Intel processors allowing more stretching, whereas AMDs will not allow so much as they invariably work to the fullest of their capabilities.