How does memory?




Internal memory (RAM, RAM), the most famous of all the considered earlier forms of computer memory. This memory is called memory “random access» («random access»), since you can access any of its cell itself. It’s enough to know the row and column, which is located at the intersection of the desired cells. There are two main types of memory: dynamic and static. Today, we consider in detail the principle of “leaky bucket”, which is based on dynamic memory. Some attention will be paid to the static memory, fast, but expensive.

The memory cell is like a leaky bucket

Оперативная память

Quite different working serial access memory (SAM). As its name implies, access to the cells of the memory is accessed sequentially. In this it resembles the film in the cassette tape. When the data are sought in such memory is checked each cell as long until you find the desired information. This type of memory is used to implement buffers, in particular texture buffer cards. That is, the SAM is used to sense when the data will be arranged in the order in which they will be used.

Similarly detail previously discussed microprocessor chip is an integrated circuit memory (IC, IC), collected from the millions of transistors and capacitors. One of the most common types of random access memory is a DRAM (dynamic random access memory, dynamic random access memory). It transistor and coupled capacitor and they form a cell containing one bit of information. The capacitor comprises a single bit of information, i.e. “0” or “1”. The transistor also plays a role in this pair of switches (switch), allowing the control circuit of the memory chip read the capacitor or change the state.

The capacitor can be thought of as a small, leaky, “bucket”, which, if necessary filled with electrons. If it is filled with electrons, the status of one. If emptied, then zero. The problem is the condenser drain. In a matter of milliseconds (thousandths of a second) full capacitor becomes empty. This means that either the CPU or memory controller to constantly recharge each of the capacitors, keeping it full. Charging should be done before the capacitor is discharged. For this purpose, the memory controller reads the memory and then writes the new data into it. This action updates the state of the memory is automatically thousands of times per second only.

Capacitor DRAM can be compared with a leaking bucket. If it is not filled with electrons again and again, his condition becomes zero. It is this update operation and made a name for this kind of memory the word “dynamic”. This memory or updated dynamically, or “forgets” all that she is “remembered.” There is at this memory drawback: the need to constantly update it takes time and slows down the memory.
Unit cell of DRAM (DRAM)

Оперативная память

Memory structure can be thought of as a three-dimensional network. Even easier: in the form of a sheet of school notebook in the box. Each cell contains one bit of data. First the column, then the data is written to specific lines via signaling on the column.

So, imagine the exercise book. Some cells are shaded in red felt-tip pen, and some were white. Red cell is a cell whose state is “1”, and the white – “0”.

Only instead of a sheet of notebook RAM is used in silicon wafer, in which “imprinted” columns (bit lines, bitlines) and rows (word line, wordlines). The intersection of the row and column is the address of the cell memory.

Dynamic random access memory transfers the charge on a specific column. This charge is called a column address strobe (CAS, Column Adress Strobe) or simply signal CAS. This signal may activate any bit column transistor. The control signal lines is called the row address strobe (RAS, Row Adress Strobe). To specify the address of a cell should be set both control signal. During recording, the capacitor is ready to accept a charge. In the process of reading the sense amplifier (sense-amplifier) ​​determines the level of charge on the capacitor. If it is above 50%, a bit is read as “1”, in other cases, “0”.

It also updates the charge cells. For the following procedure updates the counter. The time required for these operations is measured in nanoseconds (billionths of a second). If the memory chip 70-nanosecond, it means a full reading and recharging of all its cells takes 70 nanoseconds.

By themselves, the cells would be useless if there was no way to record the information in them and read it from there. Accordingly, in addition to the cells themselves, memory chip contains an array of complementary circuits. These devices perform the following functions:

Identify rows and columns (the choice of the row address and the address of the cell)
Order tracking updates (counter)
Reading and renewal of cell signal (amplifier)
Report to the cell information about whether it should hold a charge or not (activation record)

The controller has a memory and other functions. He performs a set of maintenance tasks, notably the identification of the type, speed, and memory, as well as checking for errors.
Static RAM

Оперативная память

Although static random access memory (like dynamic) is a random access memory, it is based on a fundamentally different technology. This flip-flop memory can retain every bit of information stored therein. Trigger each memory cell includes four or six transistors and includes delicate wiring. This memory is never needs updating charge. For this reason, static random access memory works much faster than dynamic. But since it contains more components of its cell is much larger than the cell heap. As a result, chip static RAM is less capacious than dynamic.

Static RAM is faster but more expensive. For this reason, static memory used in the CPU cache as well as a dynamic system RAM. For more information on static memory is written in the “Cache memory and processor register” material on overcoming the limitations of computer memory.

In today’s world, the memory chips are completed in a component called a module. Sometimes, computer scientists call it the “memory stick”. One module, or “strip” contains several memory chips. It is possible that you have heard terms such as “memory 8 × 32” or “Memory 4 × 16”. Of course, the numbers could be different. In this simple formula, the first factor is the number of chips on the module, and the second capacitance of each module. Not in megabytes, and in megabits. This means that the result of multiplication to be divided by eight to get the amount of the module of our usual megabytes.

For example, 4 × 32 indicates that the module contains four 32-megabit chip. Multiply 4 by 32, we get 128 megabits. Since we know that one byte is eight bits, we need to divide 128 by 8. In the end, we learn that “the module 4 × 32” is 16MB and is out of date at the end of the last century, that does not prevent him from being simply an excellent example for those calculations that took us.

Subject of RAM is so vast that we will return to it again. We have to find out what are the types of memory and how it is arranged module. Continued …

According to the materials
Back Disposable drones protect forests from fires
Next Robots of the Future: caretaker 30,000 laboratory mice
Tags: Memory modules , Memory , Memory .


Tags: ,

In: Technology & Gadgets Asked By: [15484 Red Star Level]

Answer this Question

You must be Logged In to post an Answer.

Not a member yet? Sign Up Now »