30.3. Memory Hierarchy

So, if you look into the memory hierarchy, so, in this case we are going to say that in the in board memory; that means, maybe it is inside the processor or it may be in a motherboard, because all indicated component of a processor or that computer will be put in a board and you say this is the motherboard. So, in board memory you are having the registers, we know these are the temporary working space inside the processor. Then next level we are having the cache memory we know that to increase the performance, to increase the speed of the processor, we used to put some high speed memory between processor and main memory. So, we say this is the cache memory and after that we are having main memory which is basically RAM random access memory, it is a semiconductor memory, and processor is going to work with the information available in the main memory.

The memory hierarchy is an essential concept in computer architecture that illustrates how different types of memory interact in terms of speed and capacity. The topmost layer is the 'in-board memory,' which includes registers located inside the processor and on the motherboard. Registers serve as fast, temporary storage for data that the processor is currently working with. Below this, we have cache memory designed to speed up access to frequently used data, acting as a buffer between the processor and the slower main memory, which is typically RAM. Main memory (RAM) is a semiconductor memory that holds data and programs that the processor is actively using, but it is slower than cache memory.

Think of your computer memory as a kitchen. The registers are like the small counter space right next to where you are cooking, perfect for keeping ingredients you are currently using. Cache memory is like a larger shelf placed near the stove, holding common spices and cooking tools that aren't in immediate use but are still easily accessible. The main memory, or RAM, is like the pantry, storing all the ingredients and tools you may need but requiring extra steps to access them while cooking.

So, for that we need some other memory element which are permanent in nature. We are going to store all the information on those particular devices and we say these are your out board storage and from that particular storage unit we are going to keep we are going to first bring the information to the main memory. So, in that particular case one is your magnetic disk or the hard disk that we used to say, most of you say that in your machine you are having either disk or hard disk of capacity say 500 GB, or 1 terabyte like that. So, this is the magnetic disk and it is started working principle is on magnetism.

In addition to temporary volatile memory like RAM, computers need permanent storage solutions for data retention when they are powered off. This layer of memory is called outboard storage, which includes devices like hard disks. Hard disks function based on magnetic principles and come in various capacities, typically ranging from 500 GB to several terabytes. This is where your operating system, applications, and files are stored permanently until they are needed by the processor.

Imagine outboard storage like a refrigerator in your kitchen. It keeps food items fresh for a long time, similar to how hard disks preserve your data. Just like how you retrieve ingredients from the fridge to cook, the processor retrieves data from the hard disk to process it. When the power is off (similar to when the fridge is unplugged), the food stays preserved inside, just as data remains intact on the hard disk.

So in this particular hierarchy what will happen? If you go down speed decreases. So, this is your low speed if you go from this top to the bottom of this pyramid. So, that registers is having a higher speed than cache memory and after that main memory like that we are having the speeds. So, if you go from the tip of the pyramid to the bottom of the pyramid that speed decreases, so this is the higher speed and here we are having the lower speed. Secondly, if you look into a capacity then capacity increases when I go from this top to bottom. So, we have very limited number of registers maybe 8 or 16, but cache memory you may have in the tune of your megabyte, if you are going to talk about main memory it is in a gigabyte. So, like that in magnetic disk now we are going to get around say terabyte capacity like that; capacity is increasing in that particular case.

The memory hierarchy is structured such that as you move from the top (registers) to the bottom (magnetic disks), the speed of access decreases while the storage capacity increases. Registers provide the fastest access speeds, but they have a limited size. Cache memory is slightly slower but offers more capacity. Main memory (RAM) is larger still, and as for magnetic disks, they offer the most storage but the slowest access speeds. This means that for a computer to operate efficiently, it relies on a combination of various memory types to balance speed and capacity needs.

Think of this hierarchy as a series of containers for storing and retrieving supplies. The registers are like small jars that hold spices—quick to access but limited in quantity. Cache memory is like a medium-sized box of commonly-used utensils—still easy to reach but holds more than the jars. The pantry, or main memory, can stock larger quantities of various foods but takes longer to search through. Finally, the basement would represent the hard disk, where you store bulk ingredients; it takes the longest to access but holds the most items.

Also if you go from this tip to bottom then if you consider about cost that cost also decreases. Since these are high speed more performance better, so we have to pay more cost. Like that if I will go there then what will happen that cost gradually decreases and we are going to get a low cost memory devices in this particular case.

In the memory hierarchy, as speed increases, the cost also increases. Therefore, high-speed memories like registers and cache tend to be more expensive per byte compared to more abundant and slower options like magnetic disks. This relationship between cost and performance means that the choice of memory types in a computer system is often a compromise between speed, capacity, and cost, with designers needing to balance these factors based on intended usage.

Imagine buying different types of storage containers for your kitchen. A high-end, insulated glass canister that keeps sugar fresh is much more expensive than a simple plastic bin meant to hold flour. While the glass canister keeps its content in the best condition (like fast memory enhancing performance), the plastic bin holds more at a much lower cost—even if its preservation capability isn't the best.

--