Find a Question:
Checking RAM for errors
We continue our earlier started acquainted with the most important component of any computer and digital devices like him: RAM. In the hierarchy of computer memory it occupies one of the upper stages. Today it is hard to believe that at the dawn of the microcomputer era operative memory is not standardized and was proprietary. In other words, the developers computers provide their devices are not similar to other memory modules. But time does not stand still and it is time to first standard linear module SIMM (single in-line memory module, the memory module with the one-sided arrangement of chips). Sometimes memory is also called “bars”. On them, as well as methods for errors, today we’ll talk.
This module has a 30-pin (30-pin) bracket (connector). Its size was approximately 9 x 2 inches. In the computer was set to two SIMM-module, followed in most cases identical to choose the speed and volume. This was due to limitations of the system bus. Let’s say you want to install two 8-megabyte memory module until the total amount of RAM of 16 MB.
Each SIMM-module can send 8 bits of data at a time. This bus system may skip both 16 bits. Later SIMM modules are somewhat larger: about 11 x 2.5 inches. Modules have found a new type of 72-pin connectors, so that was expanded bandwidth and an opportunity to support the RAM to 256 MB. At the time, this amount of memory seemed truly gigantic.
The speed and bandwidth of processors and memory had grown to meet the growing needs for performance computers. The answer to the dictates of time have become a two-way memory chips located DIMM (dual in-line memory module). They had a 168 – or 184-pin connector and measuring about 14 x 2.5 inches.
The volume of the DIMM ranged from 8 megabytes to 1 gigabyte. They have not necessarily been put in pairs, you can get by with one strap. The fact that the DIMM-module circuits which are arranged at both sides of the strip, can be considered as two merged together modules SIMM. Most of the memory modules to PC, as well as Mac G5 was 2.5 volt. But Mac G4 computers used 3.3-volt strips of RAM. RIMM memory size and configuration pins (pins) corresponds to the DIMM, but it is a special memory interface significantly improves performance. More details about the RIMM memory and other types of random access memory, we have already spoken.
With laptops history to repeat itself. Many manufacturers in the early preference for proprietary modules. But over time, the spread began to receive standard SODIMM (small outline dual in-line memory module), which is a miniature version of the DIMM. Its dimensions are approximately 5 x 2.5 inches. SODIMM have a 114 – or 200-pin connector. In subnotebooks have found application modules MicroDIMM with 144 – or 172-pin connector. In size they have less of the “notebook” slats.
The parity check method for memory errors
Modern memory modules are quite reliable components. In most modules, there is a memory controller, which at the beginning of the device checks for errors. In a chip with a built-in error checking is most commonly used method called parity. For every 8 bits of data parity chip offers extra bit.
When the transfer of one byte (8 bits) of data parity bit acquires a certain value. Chip adds the values of bits in a byte if the addition result is expressed by an odd number, the parity bit value is equal to unity. If even, then zero. When data is read again, the values are added again and compared with the value of the parity bit. For example, if the sum is an odd number and the value of the parity bit “1”, then the data is considered valid and sent to the CPU.
If the sum is odd, the parity bit and took a zero value, the chip “knows” that somewhere in the eight bits of the error and uploading data. There is a so-called “dump”. The opposite situation: the sum of an even number and the value of the parity bit – a unit. In this case, the data are also considered invalid.
At parity method has a significant drawback. With it, the error can be detected, but nothing can be done to correct them. If the data byte does not match its parity bit, data is simply discarded and the system attempts to repeat the action. Increased requirements for fault-tolerant computer operating in sensitive areas.
ECC. Memory, correcting errors
In important servers used a slightly different method of working on the bugs: correcting code ECC (error-correction code, error correction code). Like the parity in the error correction method, extra bits are used to verify each byte of data information. The difference is that each byte has not one but several parity bits. The number of check bits is dependent on the width of the tire.
In the ECC-memory involved a special algorithm to not only detect the error in a single bit, but to correct it. ECC-memory is able to identify such cases and when a single bit is wrong is not one bit, but several. However, such failures are rare, and, unfortunately, can not be correct, even with the use of ECC-method.
According to the materials computer
Back Home 3D-printing is injurious to health
Next inconspicuous wearable electronics – that’s the future
Tags: Memory modules , Memory , Memory .
Answer this Question
You must be Logged In to post an Answer.
Not a member yet? Sign Up Now »
Star Points Scale
Earn points for Asking and Answering Questions!