Many times you will find that the performance of a processor , or simply its power, is expressed in a unit of measurement called MIPS . In this article we are going to tell you what this unit of measurement is, how it is calculated and how you should interpret it to know the performance of a CPU.
In the old days, the name MIPS was used for a type of processor manufactured by MIPS Technologies, whose name came from the initials “Microprocessor without Interlocked Pipeline Stages” and that had RISC architecture, used in Sony PlayStation, PlayStation 2 and PSP consoles. But here we are not referring to that type of MIPS, already in disuse, but to the power measurement unit of a processor. Let’s see it.
What are MIPS?
In this area, the MIPS also come from an acronym in English that in this case would be “Millions of Instructions Per Second”, or millions of instructions per second . As you will already assume they are a way to measure the power of a processor, since the more millions of instructions per second the processor is capable of, the more powerful they will be.
However, this measure is only useful for processors with the same set of instructions and as long as benchmarks compiled with the same compiler and with the same level of optimization are used. This is because the same task may require a different number of instructions if they do not have the same sets of instructions. In other words, you can measure the power that a processor has to decompress files, but if the processors do not have the same set of instructions, one will use different instructions than the other and therefore the result would not be comparable.
Likewise, depending on the benchmark in which the MIPS are measured, one result or another will be obtained. Continuing with the example of compressed files, the same result will not be obtained using 7Zip as WinZIP since they use different compression algorithms and even if two processors with the same sets of instructions are compared, the results will vary greatly.
How is this measure used and what is it equivalent to?
Ok, we already know that MIPS are the millions of instructions per second that a processor can process, but this will depend on the set of instructions and the benchmark that is used. Still, it is a measure of power and if we compare two processors under the same circumstances it is perfectly valid.
However, it should be borne in mind that in this case the RAM of the system also influences a lot (therefore, if the exact same processor is used, benchmarks such as the 7Zip are often used to measure the performance of the RAM more than the processor), and is something that is not usually considered when calculating it. Due to these problems, measuring the power of a processor in MIPS has fallen into disuse since it is extremely difficult to obtain reliable measurements.
In any case, it is worth mentioning that in GNU / Linux environments MIPS usually refer to BogoMips , whereas if we compare this unit of measurement with floating point arithmetic, we should do it with FLOPS .
Therefore, in the past this was a much more valid unit of measurement than today; For reference, the first processor considered as a user, the Intel 8086, had 800 KIPS (thousands of instructions, it did not reach millions), while a Pentium 4 managed to reach 1700 MIPS.