Where is the definition of Gada Electronics

Chipset

The central element on a motherboard is the chipset. Although the main processor is the key element in a computer, it is the chipset that ensures that the various components can communicate and work with one another. The chipset is the link between the individual components of a computer. No matter what happens in a computer, the chipset is always involved.

The chipset is usually just a semiconductor component that combines many controllers. The controllers ensure the connection of the main processor (CPU), drives, expansion cards and other devices via interfaces.
The chipset ensures that the controllers are linked to one another and connected to the main processor. Different voltage levels, clock frequencies and protocols are taken into account or converted to one another.

The chipset has a major impact on the overall performance of the computer system. It controls the interaction and the flow of data between the processor, the main memory, the bus systems and the controllers of the internal and external interfaces. In general, the chipsets from the different manufacturers can have performance differences of up to 10%. If the chipset or motherboard manufacturer has made savings in development, this can lead to significant performance losses for the entire system. What a fast processor or memory cannot compensate for.
The chipset is configured via the settings in the BIOS. Depending on the chipset, RAM and processor, different settings can trigger further significant differences in performance.

There are chipsets from processor manufacturers who want to create the best possible conditions for their processors. They want to offer their processors a work environment in which their performance can be fully demonstrated. Other chipset manufacturers usually operate in the low-cost area or develop chipsets for special applications.
Choosing the right chipset is important because it determines what components can be used in a computer system. Which components the chipset supports depends on the manufacturer. The motherboard manufacturer can also install components that are missing in the chipset on the motherboard. For example for USB, SATA, Ethernet, HD Audio, I2C (SMBus), SPI and Low-Pin-Count (LPC) for a super I / O chip that supports old interfaces. If that is not enough, a system can only be upgraded with expansion cards.

The connection between the main processor and the chipset is often a modified PCI Express (PCIe). The central part of a chipset is therefore a PCIe switch, which leads out some PCIe ports on the motherboard for expansion cards and at the same time connects the controllers.

Chipset architecture

The chipset consists of several individual chips that connect the various elements of a computer on a logical and physical level.
The classic chipset architecture roughly consists of two parts. One part takes care of the communication between the main processor (CPU), memory and graphics card. The second part connects all other components into the system via interfaces. For a long time, both chipset parts were divided into two semiconductor modules.

Various chipset architectures have emerged over the course of time. The changes to the architecture primarily resulted in an increase in system speed. Inevitably, more and more functions and interfaces migrated to the processor. Modern processors are not only characterized by high computing power, but also by a high density of interfaces.

Chipset: "TTL Grab"

At the time of the 386 processor, the chipset consisted of many individual components. The name "chipset" also comes from that time. At that time, chips in dual-inline packages (DIL) were mainly used, each of which contained only a few functions. Due to the amount of chips, based on the TTL circuit family, the name TTL-Grab arose. Because the chips were arranged close together on the motherboard. Hard disk host adapters and external interfaces were designed as separate expansion cards.
With increasing integration density, the importance of bus systems for expansion cards decreased. Interfaces for mass and removable storage, network cards, sound cards and graphics cards are integrated in the chipset or even in the processor and a connection is made directly on the motherboard. The advances in semiconductor technology led to the integration of more and more functions in a single component. Over time, the chipset has developed from many individual components to a few highly integrated circuits.

Chipset: Bridge Architecture (Northbridge and Southbridge)

With increasing integration density, important and frequently used interfaces have been integrated into the chipset. This is how the division into two chips came about. If you see the basic interconnection of the chipset as a map, the northbridge is in the north and the southbridge in the south. The names are originally from Intel.

Chipset: hub architecture

With the hub architecture, Intel wanted to introduce a new generation of PCs. The reason for a new architecture was the slow connection between northbridge and southbridge. The slow PCI bus had become the bottleneck. The ever faster working processors were slowed down by the connection of the periphery. The data could not be transferred quickly enough within the computer system. To solve the problem, Intel, as a chipset manufacturer, wanted to influence the market in order to offer its own processors a better working environment.
Because Intel was the first to be active with the hub architecture, the hub architecture from Intel is described here as an example. The chipsets from other manufacturers have a similar structure. Only the functional units have different names.

Although the hub architecture is similar to the bridge architecture, some things have changed. The division of at least two chips, as well as the responsibility have remained essentially the same. The memory controller hub (MCH) functions as a northbridge. This is where the processor, memory and graphics card come together. Instead of the MCH, there is also a Graphics Memory Controller Hub (GMCH). Here the function of the graphics card is integrated into the chipset.
The internal connection between Northbridge and Southbridge is no longer the slow PCI bus, but a hub interface. It is a point-to-point connection. With this interface, further controller hubs can be connected to one another. For example, a PCI hub can be connected to the MCH, which provides slots for the PCI bus. Newer MCHs also had direct connections to PCI Express slots.
In the hub architecture, the southbridge is an extension of the northbridge in order to connect all internal and external interfaces and bus systems to the northbridge more quickly. The PCI bus and the EIDE / ATA interfaces are brought out directly from the ICH. ISA slots must be attached using an additional chip (PCI-to-ISA bridge).
When introducing the hub architecture, Intel wanted to do without old interfaces. But for reasons of compatibility, they still had to be available. There was a super I / O module for this, which was connected to the I / O controller hub via the LPC.
The LPC (Low Pin Count Interface) is a stripped-down ISA bus. LPC is used to connect the Super I / O controller or the Trusted Platform Module (TPM). The super I / O controller provides PS / 2 keyboard, COM, infrared and printer interfaces. Traditional computer end devices such as keyboard, mouse, printer and floppy disk drives are connected to it. These devices are also known as legacy devices, for which almost all operating systems have standard drivers.
Although the hub architecture is modular, the ICH and memory are connected via the SMB. There are also direct interrupt lines between the processor and the I / O controller hub (ICH).

Chipset: Advanced hub architecture

The chipset is slowly becoming less important in the expanded hub architecture. The connection of graphics card and main memory (RAM) are always the bottleneck of a computer. Fast access to the main memory in particular decides whether a processor can fully exploit its capabilities. Therefore, one of the measures was to integrate the memory controller, actually the task of the chipset, directly into the processor. In addition to the connection to the chipset, the connection to the main memory is also made directly from the processor. In this way, the processor saves the detour via the chipset when it accesses the memory. The advantage of having access to the main memory is shortened by a few cycles.
Depending on the chipset and manufacturer, the split into two chips is dispensed with and all interfaces and their control are integrated in a single module.
The increasing importance of the graphics performance of a computer has the effect that the graphics card is also connected directly to the processor.

With the integration of the memory controller and the graphics chip into the processor, the northbridge becomes less important. The CPU takes on more and more tasks. Parts of the chipset migrate to the CPU. These include graphics cards, memory controllers and direct connection of expansion cards, especially external graphics cards and drives. Only simple interfaces are provided via an I / O chip.

Overview: computers

Overview: computer architecture

Other related topics:

NEW 5th edition of the computer technology primer

The 5th edition of the Computertechnik-Primer has been completely revised and is available as a printed book, as an eBook in PDF format and in the Google Store, for Amazon Kindle and Apple iBook.

The Computertechnik-Fibel is a book about the basics of computer science, processor technology, semiconductor memory, interfaces, drives and important hardware components.

Order or download now

Everything you need to know about computer technology.

Computer technology primer

The computer technology primer is a book about the basics of computer technology, processor technology, semiconductor memory, interfaces, data storage devices, drives and important hardware components.

I want that!