The computing nodes are mainly responsible for the performance of the computer - the central processor and the video card, it is they who are overclocked in the first place. By increasing the frequency potential, the user increases the number of operations that the computing unit can perform in one second. But besides data processing, they need to be stored somewhere else and transported at a sufficiently high speed, to match the speed of their processing. So that the CPU and GPU do not idle in vain, they also overclock the RAM, i.e. increase its throughput.
Instructions
Step 1
This is done in two ways:
The first way is to decrease the timings, i.e. memory delays when switching from one mode of operation to another. For example, in order to clear a memory cell and then re-write any data into it, it will take several clock cycles, but it should be remembered that with a decrease in timings, the frequency potential of the memory also decreases.
Step 2
Another way is to increase the frequency, everything is subject to the same law as in the first case. On the one hand, the frequency of data exchange increases, on the other hand, it is necessary to increase the delays to maintain system stability. Also giving stability, it is possible to raise the voltage supply of memory modules, but this should be done with great caution, since incorrect values or poor cooling can lead to devices overheating and failure.
Step 3
There is no definite answer which method is better. It is necessary to experimentally select the optimal combination of both parameters, so that lower timings correspond to the highest frequency. Also, special attention should be paid to cooling, since during overclocking the devices work at the limit of their capabilities and emit much more heat than with the parameters declared by the manufacturer.
