We investigated early digital computer memory, see computer history – Core memory, and stated that the current standard RAM (random access memory) chip memory. This is in line with Moore's often cited application (Gordon Moore is one of Intel's founders). It states that the component density of integrated circuits, which can be paraphrased with unit cost performance, doubles every 18 months. Early core memory in microseconds was cycle time, today we are talking about nanoseconds.
You may be familiar with the term cache as for PCs. This is one of the performance features discussed on the latest processor or hard drive. The processor may have L1 or L2 cache and disk size of different sizes. Some programs also have a cache, also called buffer, for example when writing data to a CD writer. Early CD burner programs overwhelmed the & # 39; The result of these was a good backing!
The mainframe systems used cache for many years. The concept became popular in the 1970s to speed up memory access time. This was the era when magmemory was gradually removed and replaced with integrated circuits or chips. Although the chips were much more efficient in terms of physical space, they caused other problems in relative heat and heat production. The chips of the plans were faster, warmer and more expensive than other design chips that were cheaper but slower. Speed was always one of the most important factors in computer sales, and design engineers always looked for ways to improve performance.
The cache memory concept is based on the fact that a computer itself is a sequential processor. Of course, one of the great advantages of a computer program is that it can "branch" or jump " is a subject of another article in this series. However, there is still enough time for an instruction to follow the other to make a buffer or cache useful for the computer.
The caching idea is to predict what data should be processed from memory in the CPU. Take a program consisting of a series of instructions, each stored in memory, say 100 upwards. Command 100 is read from the memory and executed by the CPU, and then the next instruction is read from location 101 and executed, followed by 102, 103, and so on.
If the memory in question is the central memory, 1 microgram for reading an instruction. If the processor takes, say, 100 nanosecons to execute the instruction, it has to wait 900 nanosecons for the next instruction (1 micron = 1000 nanoseconds). The CPU has an effective repeat speed of 1 microsecond. (Times and times are typical, but do not refer to any special hardware, but only to illustrate the principles).