Modern Motherboards and RAM

A motherboard is a printed circuit board used in a personal computer. A typical motherboard provides attachment points for one or more of the following: CPU, graphics card, sound card, hard disk controller, memory (RAM), and external peripheral devices. The most important component on a motherboard is the chipset.

Chipset
The chipset consists of two component chips known as the Northbridge and the Southbridge. The Northbridge typically handles communications between the CPU, RAM, AGP, PCI, or PCI Express, and the Southbridge. The Northbridge is the most important factor in determining the speed of the CPU and the amount, speed, and type of RAM that can be used. The Southbridge is not directly connected to the CPU. The Southbridge contains support for various buses, IDE/SATA controller, Real Time Clock, Ethernet, RAID, USB, firewire, etc.

Sockets
The CPU socket or slot is the connector that interfaces between a computer’s motherboard and the processor itself. There are different sockets for different processors. Newer sockets are named after the number of pins they contain. Older sockets were named in the order of their invention. The fit between the socket and the CPU is determined by the pin-grid array (PGA). In the early days of the PC, all processors had the same set of pins that would connect it to the motherboard. This was called a Socket 7. Any processor would fit into any motherboard. Today, processor manufacturers Intel and AMD use many different PGAs. No processors fit into the Socket 7 anymore. As new bleeding-edge processors make their way into the market, they need more pins to provide power and functionality. Thus, motherboards must advance with processors. Newer Intel CPUs do not use PGAs. Instead, the pins are part of the socket, not the CPU. This is known as the Socket T. AMD currently uses Socket 754, 939, and 940. AMD will be releasing Sockets F, S, and M2 sometime in 2006. These sockets will boast 940 and 1,207 pins.

Peripheral Slots
Peripheral Component Interconnect (PCI) provides connections for video, sound and video capture cards, as well as network cards. Accelerated Graphics Port (AGP) is another video connection. Universal Serial Bus (USB) allows the connection of many different types of devices from printers, to scanners, to mice. Ethernet connectivity allows connections to high-speed internet. Integrated Drive Electronics (IDE) provide interfaces for the hard drives. New motherboards incorporate PCI Express and RAID technologies.

Form Factor
The shape and layout of a motherboard is called the form factor. The form factor determines the shape of the computer case as well as where components will go. Differences between form factors can include; physical size and shape, mounting hole location, feature placement, power supply connectors, and others. The most common form factors is the BTX, developed by Intel.

Motherboards of the Future
Both AMD and Intel have developed 64-bit processors, which means that new motherboards have been designed and created to support the doubling of processor speeds. There are rumors that companies are beginning to design hybrid motherboards that would allow either an Intel or AMD processor. These ideas are currently in development, but there are questions about its success.

What is RAM?
RAM is short for random access memory. It is a type of computer storage whose contents can be accessed in any (hence the random) order. RAM is typically used for primary storage to hold information that is being actively used and manipulated.

Early “RAM”
Early computers used vacuum-tubes, which basically behaved like the RAM we use today. Wires accessed information that was stored in magnetic cores based on the polarity of the magnetic field.

Types of RAM
Static RAM (SRAM) retains its contents as long as the computer has power. Dynamic RAM (DRAM) must be refreshed periodically. Both are volatile, which means when the computer loses power, the data is lost.

How DRAM Works
A memory chip is an integrated circuit which is made up of millions of transistors and capacitors. Think of a capacitor as a bucket that can hold a certain number of electrons. To store a 1 in a particular memory cell, the bucket is filled. To store a 0, the bucket is emptied. In dynamic RAM, the bucket has a leak. The memory controller and CPU must keep refilling the bucket (refreshing the memory). Memory cells are engraved in an array of rows and columns. The intersection of a particular row and column is the memory address. A charge is sent through the column to activate the transistor at each bit in the column. Other circuits perform functions such as locating the memory address, keeping track of how many times the memory has been refreshed, etc.

How SRAM Works
A form of “flip-flop” holds each bit of memory. A flip-flop for a static memory cell takes four or six transistors. The storage cell that the transistors form has two “stable states”. One state is a 0, the other is a 1.

Types of SRAM & DRAM
XDR DRAM is a high performance RAM interface that is effective in small consumer systems such as Sony’s new PlayStation 3. Synchronous dynamic random access memory (SDRAM) is the most common form of RAM in desktops today, the key to this is pipelining which means the chip can accept a new command before finishing the previous one. Double data rate synchronous dynamic RAM (DDR SDRAM) is just like SDRAM except that is has higher bandwidth, which translates to greater speed. Rambus dynamic random access memory (RDRAM) uses a special high speed data bus different from DRAM architecture. It was phased out in 2003 and was primarily used in Sony’s PlayStation 2. Double Data Rate Two Synchronous Dynamic Random Access Memory (DDR2 SDRAM) is becoming the mainstream standard for personal computer memory. It can run at higher speeds due to an improved electrical interface (lower power usage). Already in development is DDR3, which will run at speeds even faster than DDR2.

RAM Packaging
RAM packaging refers to the board and connector used for RAM. Early variants of the single in-line memory module (SIMM) had 30 pins and provided 8 bits of data, the second variant had 72 pins and provided 32 bits of data. The dual in-line memory module (DIMM) had 72, 144, 168, 184, or 240 pin connections, replaced SIMM, and ranges in capacity from 8 MB to 1024 MB (1 GB). Small outline DIMM (SO-DIMM) is a smaller version of the DIMM that is used in laptops and has 72, 144, or 200 pins.

Buffered vs. Unbuffered
Buffers improve memory operation. A buffer is added between the memory module and the memory controller. There is no loss in performance with buffered memory. Buffered and unbuffered RAM cannot be mixed! The notches/keys will show whether the module is buffered or unbuffered.

Registered vs. Unregistered
Registering is similar to buffering in that it also improves memory operation. The concept of registering is similar to buffering in that it places less of a load on the memory controller. Registered modules are slightly slower than unregistered ones because the process of registering takes time. Registered modules are also more expensive.

ECC and Parity
Memory chips with built-in error-checking typically use a method known as parity to check for errors. Parity chips have an extra bit for every 8 bits of data. For example, 32-bit memory becomes 36-bit memory with parity. The concept of parity is complicated. It will detect errors, but won’t do anything with them. Error correction code (ECC) operates the same way as parity in that it adds an extra bit, but it uses a different system that allows it to detect and correct single bit errors.

Memory Speed
When the CPU needs something from memory, it sends out a request to the memory controller. The memory controller sends the request to memory and alerts the CPU when it can read the information. The speed of this cycle determines the speed of the memory. It is measured in Megahertz (MHz). Speed is increasing rapidly in short amounts of time. In 1997, the average speed of memory was 66 MHz. In 2005, the average speed of memory was 667 to 800 MHz (DDR2).