Buffered memory is a type of computer memory. It is designed to control the amount of electrical current which goes to and from the memory chips at any one time. This makes for more stable memory, but increases the cost and slows the speed at which it works.
In a buffered memory system, a hardware register is located between the part of the computer which controls memory and the memory chips themselves. This is a device which can hold a certain amount of information at once. The register will fill up completely and then pass on all this information at once.
The use of a hardware register means that the computer can much more reliably control how much data transfers to and from the memory chips at any one time. The reason for doing so is that this also controls how much electrical current passed to and from the chips. This maximizes the amount of data the computer can handle at once without running the risk of overloading the memory chips with current and causing either a delay or damaging the chips.
The drawback of this system is that there is a delay while the register fills up with data. The delay is a very short amount of time and in most PCs will be one sixty-six-millionth of a second. However, over time this adds up and makes a noticeable difference in performance.
Because of the performance loss and the increased price of the hardware register to the system, buffered memory is not normally used in personal computers. Instead it is most common in servers, particularly those dedicated to data processing. In these applications, the benefits of reliability outweigh the drawbacks of cost and confusion.
Today, buffered memory is more commonly referred to as registered memory. The terms are generally interchangeable. However, you may also see the phrase “fully buffered memory.” This is slightly different as it involves every part of the data being buffered, while in standard buffered or registered memory, only certain parts are buffered.
Buffered memory should not be confused with buffer memory. This is a term sometimes used to refer to cache memory, such as a disk cache. This involves using part of a hard drive to store information which is likely to be used repeatedly during a computing session, which saves time that would otherwise be spent repeatedly accessing it from the original source. Strictly speaking, “buffer memory” should not be used to refer to cache memory, as “buffer” means a location where information is only intended to be stored and accessed once.