The unknowing user can find two identical processors in terms of basic specifications. Same number of cores, same clock speed measured in MHz or GHz, that’s when, given the huge number of models on the market, it’s lost for lack of understanding the information.
When is a processor faster?
When several factors are given, but above all there are three: architecture, clock speed and number of cores. Although in general the former is generally ignored due to the fact that it is not quantifiable. Which leads to confusion about the speed of a processor.
We tend to think of the speed of a processor as marked by the clock speed, whereas what is marked is the number of pulses produced per second. Today, a CPU and a GPU work in stages to execute a program instruction. Also, depending on the instruction type, the number of steps varies. Think of it in a factory where every part of the assembly line moves at the tick of a clock.
That is, when the first instruction is in the second step, the second enters the first and so on until the first is resolved. We seek to measure the average number of instructions carried out per clock cycle. The fastest processor will be the one with the highest number of all. Obviously having a better clock speed and more core count helps. Although, in the same way that not all 110hp cars are that fast, not all 3GHz processors are that fast.
Neck
However, it must be taken into account that these measurements are not always completely accurate, since we can find the following three problems, which affect the speed at which the processor will go.
- The units in charge of solving certain instructions are totally or partially occupied.
- The data required by one or more instructions is not in the nearest registers or caches. So there is an additional delay.
- The program has a specific selection of instructions that harms or benefits that processor model.
The most problematic bottleneck is memory, since the speed of it did not evolve in the same way, the cache memory patch had to be extracted at different levels. Initially it was a single tier, but as the latency between CPU and RAM increased, new tiers were added. A few years ago it was unthinkable for a processor to have 3 levels of cache, today it is a general rule, even in graphics chips or GPUs. Unsurprisingly, AMD’s RDNA 2 architecture on PC already uses L3 cache.
Another problem is that of contention, this occurs when two instructions that operate in parallel must share the same processor resources. When designing a processor in order to save space and transistors, it happens that some instructions share resources and combine them, assuming at the same time that they are solved in more cycles than usual . When designing a processor, engineers ensure that this problem does not occur with the most common instruction combinations.
