Categories: PCs and Hardware Glossaries Solutions: Printer Memory Versus System Memory Memory's Too Fast? L2 Cache Specifications How Computers Use Memory More Hardware Solutions

By Neil Randall  

Once upon a time, magazine ads for computers were actually readable. Not only were there fewer processors to choose from, there were also fewer choices in everything else, from hard disks to sound cards. Today, the choices boggle the mind. Acronyms run rampant. And as if it weren't bad enough having to deal with all the other jargon, we now have to confront a seemingly limitless variety of memory types. It used to be enough to have RAM, but now RAM always seems to have one or more extra letters in front of it. DRAM, VRAM, SRAM, SDRAM, WRAM--it's enough to confuse even the staunchest technology watcher.

In this article we hope to ease some of the confusion by presenting a description of the various types of RAM. We don't have the space to cover them all, but you'll find the most frequently encountered RAM varieties in this glossary.

The first order of business, however, is a short explanation of how RAM works, as well as the relationship between RAM and the CPU (central processing unit, or processor). The CPU--yours is probably an Intel 486 or Pentium type--is the heart of the computer, where data is processed and program instructions are interpreted. Integrated with the CPU is the system's main memory, called random-access memory or RAM. Together, these two components make up the core of your machine; components such as hard disks, controllers, and video cards are peripheral to this central activity, and are therefore known as peripherals.

The CPU uses its RAM as a storage area for data, calculation results, and program instructions, drawing on this storage as necessary to perform the tasks required by programs. In order to store data and draw from the data store, the CPU specifies the memory address of the required information. The address bus allows the CPU to send the address to RAM, and the data bus allows the actual data transfer to the CPU. The term bus itself refers to the connection between the two devices that allows them to communicate. An important measurement of RAM performance is access time, the amount of time that passes between the instant the CPU issues an instruction to RAM to read a particular piece of data from a particular address and the moment the CPU actually receives the data. Today's RAM chips typically have a 60-ns access time, which means it takes 60 nanoseconds (a nanosecond is a billionth of a second) to perform this round-trip function. This access time is much faster than that of the 100- to 120-ns chips of a few years ago, but it's still much slower than the ideal access time of zero, which would be realizable if the CPU itself stored all the data. To speed things further, the CPU has access to cache memory (usually referred to as "the cache"). At 20 ns or better, cache memory is faster than main memory, but systems contain less of it than main memory (cache memory is expensive), and therefore only select data--the data the CPU will probably need next--is placed inside it. The selection is handled by the cache controller.

Memory chips function by storing electronic charges. The chips are made up of a capacitor and a transistor, with the capacitor storing the charge and the transistor turning the charge on or off. With RAM chips, the system can alter the on or off state of the charges, but with ROM (read-only memory) chips the charges are either permanently on or permanently off. This article deals only with RAM.

All RAM technologies emphasize speed and attempt to offer more of it without an increase in cost. But CPU technology keeps getting faster, and memory technology must keep pace--hence the need for different types of RAM. It may be confusing, but rest assured, it's in our best interests.

Continues...    

Published as Tutor in the 10/21/97 issue of PC Magazine.