RAM temporarily stores programs and data while they are in use. RAM is known as volatile memory, because the data stored in RAM does not survive a power outage. Examples of non-volatile memory are floppy disks, hard drives, and tape drive tapes; data stored on these media survives power outages.
RAM provides the working area for the PC to perform its tasks. RAM can be compared to the desktop — a limited space where work is performed — in an office. The file cabinet is the non-volatile memory used for storage of data not in use.
The two types of RAM used in PCs today are static RAM (SRAM) and dynamic RAM (DRAM). The biggest difference between the two is the way the information in each is refreshed. With these two types of memory, several different package techniques are used that have different voltage requirements and parity characteristics.
Packaging
The physical layout of RAM and the way it is purchased is known as the packaging. When IBM-compatible PCs were first introduced, all memory came in a Dual In-line Package (DIP) and was arranged in banks. Extra memory was added in banks; these banks provided a way for the motherboard to divide RAM into usable groupings. When installing more DIP-based memory, you must install nine chips — 8 bits for a byte, plus one bit for parity. We discuss parity in more detail later in this chapter.
To make installation easier, newer systems use 30-pin or 72-pin Single In-line Memory Modules (SIMMs). A SIMM is a small circuit board with an edge connector that fits into a socket on the motherboard. The RAM chips are soldered to the small circuit board so that only the board is installed rather than nine individual chips of DIP memory. Some motherboard manufacturers use Single In-line Pinned Packages (SIPPs), which have pins instead of an edge connector for insertion into the motherboard.
Most of the latest boards on the market use a 168-pin Dual In-line Memory Module (DIMM) rather than DIP, SIPP, or SIMM.
With all RAM, numbers describe how much RAM is available with each particular chip or memory module. Older DIP memory typically was sold in 64K × 1 or 256K × 1 increments, which is read Rows X Bits. A 64K × 1 chip could store one bit of data in 64K (64,000) different rows.
To have 64K of RAM, you need nine 64K × 1 RAM chips to provide one byte (8 bits) of wide data storage and 64K (64,000) storage addresses. Similarly, the 256K × 1 chips are installed nine at a time to provide 256K (256,000) byte-size storage locations. The ninth chip does not provide more storage capacity; it is used for parity (we discuss parity in more detail below).
With SIMM, SIPP, and DIMM packaging, the same rules apply, but each module is 8 or 9 bits wide, so the numbering scheme simply assumes this to be true. These modules are difficult to size by looking at the number of chips on each module, because they don’t indicate how deep (how many addresses) there are.
On example, on the 30-pin SIMM module, each module can have 3, 8, or 9 chips. The 3-chip style uses two 4-bit-wide chips and one 1-bit-wide chip to provide 8 data bits with parity. The 8-chip style uses eight 1-bit chips for data and provides no parity bit. The 9-chip style provides nine 1-bit chips; eight are for data and the last one is for parity.
On 72-pin SIMMs and 168-pin DIMMs, the size is easier to discern by the description, but no easier to identify by looking at the component itself. To calculate the storage space available, multiply the first number (in the description) by the second number, then divide by 8. For example, using the formula, the 4 × 64 DIMM translates in this way: 4 multiplied by 64 equals 256; 256 divided by 8 is 32; therefore, you have a 32 MB DIMM module.
Banks
The ability of a PC to access more than one row of RAM at a time is known as banking. On the 8088 processor chip, the external data bus is eight bits wide; therefore, memory access needs to be only eight bits wide. The 80286 provides a 16-bit-wide external data bus; therefore, it makes sense to access the memory two bytes (16 bits) at a time. To do this, the PC uses banking. Figure 2.4 shows how banking works.
The number of SIMMs needed to create a bank depends on the processor type used, because different processors have different data bus widths. Use the following list to determine how many SIMMs it takes to create a bank for the various processors:
8088 — N/A
286/386SX — 16-bit external data bus, therefore two 30-pin SIMMs create a bank
386DX/486 — 32-bit external data bus, therefore four 30-pin SIMMs create a bank
Pentium — 64-bit external data bus, therefore eight 30-pin SIMMs create a bank
Note: Because 72-pin SIMMs are twice as wide as 30-pin SIMs, it takes half as many 72-pin SIMs per bank. Because 168-pin DIMMS are four times as wide as 30-pin SIMMs, it takes one-fourth as many DIMMs as 30-pin SIMMs per bank.
Banking RAM provides the same type of efficiency as adding more lanes to a highway. With banked RAM, more data can be accessed at one time.
Parity
On PCs, most data transfers occur in increments of eight bits — a byte. When memory is manufactured, a ninth bit is added and is called the parity bit. Using the parity bit, the memory-control circuitry verifies that the data stored RAM is correct. If the memory-control circuitry detects an error, an error message is displayed. In many of the newer operating systems, this error message may cause the system to lock up. This setup is known as error-correcting memory.
The odd parity method of checking errors has become the standard used in PCs. In the odd parity method, the number of bits in a byte are added. If an even number of bits are on, then the parity bit is turned on. If an odd number of bits are on, the parity bit is left off. When the data is retrieved, the number of bits are added, and the results are checked against the parity bit. If the results agree, the data is considered good.
Some newer PCs let you disable the error-correction circuitry and use cheaper, non-parity-checking memory. When working with SIMM memory, you must make sure that the SIMM has parity if the motherboard requires it. If the option is available, you should always use error-correcting (parity) memory for better system reliability.
Dynamic RAM
Dynamic RAM (DRAM) is the most common type of RAM because it is the least expensive. DRAM must be refreshed regularly by circuitry built into the motherboard. The refresh function takes time away from the system that could be used for processing, which makes DRAM slower than Static RAM.
Static RAM
Static RAM (SRAM) is a type of RAM that does not have to be refreshed constantly. When information is placed in SRAM, that information stays until it is replaced with new information or until the system is powered off. This type of memory is extremely expensive because it requires about five times the internal circuitry as DRAM. Because of the cost, SRAM is only used for cache in PCs. Cache is discussed in detail in Chapter 5.
Extended Data Out RAM
Extended Data Out (EDO) RAM is one of the latest technologies available with 72-pin SIMM and 168-pin DIMM modules. EDO SIMM and DIMM chips can increase performance by approximately 20 percent because of their dual-pipeline architecture, which allows EDO memory to have a timing overlap between successive accesses. To use EDO memory, the motherboard chip set must be designed to support the EDO RAM.
Synchronous Dynamic RAM
Synchronous Dynamic Random Access Memory (SDRAM) is one of the latest additions to the high-speed RAM family. As the “synchronous” portion of the name suggests, this RAM is tied to the system clock, which provides more efficient RAM access. Being tied to the system clock, along with SDRAM’s ability to accept pipeline instructions, gives SDRAM the ability to perform four to six times more efficiently than standard DRAM. At present, SDRAM always uses DIMM packaging.
Video RAM
In today’s graphic-intensive operating systems and applications, the video subsystem has become a major player in the overall performance of the system. Graphic accelerator cards that greatly enhance the overall performance of the system are now available.
Video RAM (VRAM) is specially designed video memory that provides dual-ported access to the video memory. This dual-port capability makes it possible for the video board to write data to the video port while receiving data from the CPU, thereby greatly increasing the speed of the video and of the overall system.
Other Graphics-Related RAM
Windows RAM (WRAM) is another dual-ported RAM that some video cards use. WRAM is slightly more efficient than VRAM but is not as widely used and accepted.
Another type of video memory that bears mentioning is Synchronous Graphics RAM (SGRAM), which works in the same manner as SDRAM. It is extremely fast and widely used on many of today’s video cards.
Order Your SQL Fundamentals CD Today! Learn how to use SQL Server, understand Office integration techniques and dive into the essentials of SQL Express and Visual Basic with this free SQL Fundamentals CD.
You've Deployed SharePoint...Now What? This one-day free online conference delivers the technical knowledge needed to kick MOSS up a notch. In one information-packed day, independent SharePoint experts will present practical, real-world information and provide take-away, ready-to-use solutions
What Would You Do If You Ran Microsoft? ITTV's 2008 inaugural video contest, "If I Ran Microsoft..." is your chance to tell it like it is. Be goofy or be serious, but don"t miss this chance to have fun, win prizes, and go viral in a major way.
Maximize Your SharePoint Investment This web seminar discusses how true bi-directional replication of SharePoint content from one server to another enables branch offices to maintain access to current SharePoint content.
Windows IT Pro
Register
