Chipsets

Chipsets, Cache, Buses and more

The bus structures and interfaces supported by the motherboard and CPU are controlled by functions included on the chipset. The chipset is a group of chips that together help the processor and other peripheral devices plugged into the motherboard communicate with each other. The chipset controls the bits (data, instructions, and control signals) that flow between the CPU and system memory over the motherboard's buses. The chipset also manages data transfers between the CPU, memory, and peripheral devices. It also supports the expansion bus and any power management features of the system. However, the chipset contains only enough instructions to issue control commands to device drivers, which are what actually control the peripheral device.

Chipsets are integrated into the motherboard and usually cannot be upgraded without changing the entire motherboard. A PC's chipset is matched to the motherboard and the CPU as a set. Some chipsets support more than one processor, but in general, a given chipset is matched to a single processor type. Along this line, you will commonly see the chipset referred to by the CPU's mounting on the motherboard, for example, Socket 7 chipsets, Socket 370 chipsets, or Slot A chipsets.

Chipsets, especially two-chip chipsets, are divided into a North Bridge (the larger chip) and a South Bridge (the smaller chip). The North Bridge provides support and control for main memory, cache memory, and the PCI bus controllers. The South Bridge provides control for peripheral devices and those controllers that are not essential to the PC's basic functions, such as the serial port controller.

The chipsets are not the only controller sets on the motherboard. The most prominent controller sets are the keyboard controller and a superset of input/output device controllers called the Super I/O controller. The Super I/O chip combines controllers that are common to all systems. Controller chips are also found on many high-end devices and adapter cards.

Understanding the internal bus

You should have a good understanding of general bus architectures for the Core Hardware exam. You may not be asked any specific questions from this area, but it will sure help you understand some of the questions you will be asked.

Time Shaver Don't worry about the difference between the specific system bus structures of a 386 processor and those of the latest Pentium technology. Concentrate on the various bus architectures used in device I/O (chipsets and expansion cards), which is where you, the repair technician, most likely come into contact with the bus.

The internal bus, also known as the system bus, is that maze of wires on the motherboard. It provides the internal components of the computer with four different necessities:

Power: Power comes to the motherboard straight from the power supply. The motherboard uses the system bus to distribute power to components mounted on or plugged into it.

Control signals: The control unit within the CPU sends out control signals to coordinate the activities of the system. These signals are carried on the internal control bus.

Addresses: PC components pass data and instructions between one another using memory location addresses to reference the location of the data or instructions in memory. Addresses are transmitted on the internal address bus.
Data: Data and instructions are transferred between components on the internal data bus.

Caching In on a Good Thing

Expect to see questions on the A+ Core Hardware exam on cache memory, which is a motherboard component. Understand what it exists, why it is, its limitations, and how and when it is used. One of the mysteries of the PC is that nearly all of its components, including the processor, the memory, the motherboard data buses, and the hard drive, operate at different speeds and data transfer rates. You may think that these parts could be coordinated better, but because they are the products of competition, they cannot. No single company makes all of the components that go into a PC (although Intel certainly is trying). Each of the competing companies is trying to develop the fastest, biggest, and best computer component possible.

Two motherboard components that must overcome their differences and work together are the CPU (processor) and primary memory (RAM). RAM works in nanoseconds (billionths of seconds) and is seemingly faster than the CPU, which works in megahertz (millionths of seconds). However, when the CPU requests data from the RAM, it takes a fair amount of time to locate the data and then transfer it over the data bus to the CPU. No matter how fast the RAM may be, the CPU must wait while all of this is happening, and this is bad! One of the underlying design goals of the PC is to prevent the CPU from being idle as much as possible. This is where cache memory comes in.

Thanks for the cache memories

Cache memory is an extremely fast memory type that acts as a buffer between RAM and the CPU. It holds frequently requested data and instructions so that they are immediately available to the CPU when needed. Yes, it is a bit more complicated than that; but for the A+ exams, just remember that cache memory and the caching process hold data and instructions from a slower resource or process so that they are ready when needed by a faster device or process.

A cache is a buffer that is used to mitigate the speed differences between devices. Today's PCs commonly include cache memory between the RAM and the CPU and perhaps between the hard drive and RAM, as well. Caching is used in two ways on the PC:

l Cache memory: This is a relatively small and very fast memory storage located between the PC's primary memory (RAM) and its processor (CPU); it is used to hold data and instructions retrieved from RAM to provide faster access to the CPU.

l Disk cache: This cache buffer is used to speed up the transfer of data and programs from the hard drive into RAM. Disk cache, which is either in RAM or some additional memory on the disk controller, holds large blocks of frequently accessed data.

Cache memory is usually a small amount of Static Random Access Memory or SRAM. SRAM is made up of transistors that don't need to be frequently refreshed like DRAM, which is made up of capacitors. SRAM is very fast, with access speeds of 2ns (nanoseconds) or faster. This is much faster than DRAM, which has access speeds around 50ns. Because of its speed, SRAM cache memory can transfer data to the CPU at a much faster rate than it would take to transfer the same data from main memory. Another contributing factor to the speed of the cache is its proximity to the CPU, which eliminates most of the latency (delay) involved with transfers from RAM. SRAM isn't used for primary memory in a PC for very good practical and economic reasons. SRAM can cost six times more than DRAM and requires a lot more space on the motherboard to store the same amount of data as DRAM.

Cache is also commonly referred to by its level or proximity to the CPU. Cache is designated in two levels:

l Level 1 (L1) cache: Level 1 cache is often used interchangeably with internal cache and rightly so. L1 cache is placed internally on the processor chip and is, of course, the cache memory closest to the CPU.

l Remember Level 2 (L2) cache: L2 cache is normally placed on the motherboard very close to the CPU; but because it is not inside the CPU, it is designated as the second level of cache.

Although L2 cache is commonly considered the same as external cache, L2 cache can also be included on the CPU, just a little behind L1 cache.

Level 1 is not higher in ranking than Level 2 cache. The levels of cache work together, and data are located on either level, depending on the rules and policies of the caching system. Level 1 cache cannot be increased without changing the CPU. On the other hand, L2 cache can be upgraded on most motherboards. L2 cache modules plug into special cache module mounts or cache memory expansion sockets on the motherboard.