Integrated video card

Встроенная видеокарта

The built-in video card (it is often called an integrated or "onboard" video card) is part of the chipset of the system logic of the computer's motherboard (part of the chipset). The integrated video card, as a rule, is located inside the chip of its "north bridge".

Do not worry if you do not quite understand what it is all about (some "bridges", etc.) We will talk more about the motherboard device in more detail in the following sections. While we are interested only in the built-in video card. The photo shows a typical example of integrated video on the motherboard.

Integrated video card Integrated video card

The figure "1" here we have this mysterious "north bridge". As you can see, there is nothing special about it: a large microcircuit, under the lid of which is located (among other things) a built-in video card. Labeling the chipset (the inscription on it) can be completely different, depending on the manufacturer. In this case, we see that this chip was released by the firm "SIS" figures below - its serial number and model.

Note: in modern computers chipset chips are often covered with radiators to dissipate heat. So it's not so easy to look at markings.

Under the number "2" in the photo above we have a processor socket (all in holes for processor "legs"). The number "3" indicates two connectors for memory modules .

Now let's see what are the features of the built-in graphics card. Let's take a closer look at this point. What is and what does not have two different types of video cards: integrated and external (discrete)?

Here are the differences: an external card can be replaced with a new one. All modern external video cards have a powerful and productive graphics core, covered by a high-performance cooling system. They have their own (soldered on the video card itself) memory, and the memory is faster than the operating memory. Also - a high-speed data bus, the characteristics of which depend on the type of connector for connecting the video card to the motherboard (PCI, AGP, PCI-Express, etc.). Discrete solutions are also equipped with various outputs for connecting monitors and televisions.

The built-in graphics card is an integral part of the chipset chipset system logic, which can not be replaced (unless the chip itself is replaced). Embedded video, by definition, is much less productive than its discrete counterpart. The fate of such video cards is inexpensive office computers that do not require a powerful graphics processor.

The built-in video card does not have its own RAM, but uses the memory installed on the motherboard. Hence - an additional reduction in performance (data from the video card is transferred first to the memory controller, then to the computer's memory itself, and then it goes to the CPU for processing). In short, it's a long story! :) Yes, and they are transmitted not by a specialized data bus, but by a common system bus on the motherboard, which further reduces the speed of performance of such systems.

The integrated solution has one standard VGA connector on the back of the system unit for connecting a monitor or a TV (in modern models there are instances that also have a digital DVI / HDMI connector).

Built-in and external video card connectors Built-in and external video card connectors

In the photo above, under the number "1" we see a video output, which is equipped with a built-in video card. Under the number "2" - one of the video outputs of the video card is discrete (made as a separate expansion board).

As we already mentioned, the built-in video card is the destiny of office computers of medium power. You can not run a computer game on this card with realistic 3D graphics and complex effects. More precisely, you can launch it, under certain circumstances, but it will be "braked" - godless! Do you need it? It's better if you can not do without games - get yourself a good graphics accelerator dollars for 150-200 and for a few years forget about this problem :)

For example, on good motherboards, which, according to the manufacturer, will be used in multimedia gaming configurations, the built-in video cards are not even installed (and correctly, why are they needed if they are not used anyway?).

How to determine the built-in video or not? A - by eye :) Once again, look closely at the back of the system unit (photo above), notice how the connector is located under the number "1", and how - under the number "2"? Now mentally imagine that the case is transparent and "see" how the motherboard is located in it (parallel to the side cover), respectively, and the VGA output of the integrated video is the same.

Now look at the figure "2" - the connector is oriented perpendicularly to the motherboard, just as an external (discrete) video card is installed in the graphics accelerator slot.

External video card External video card

Now in the market there were central processors with integrated in them graphic core. The idea with additional functions integrated directly into the CPU (the graphics chip is located on the same chip with a processor, a memory controller and a cache, and not a separate microchip), in 2006, AMD said. A little later, she also introduced her concept of the APU (Accelerated Processing Unit) and implemented it.

Previously, this whole thing was clearly divided:

  • CPU - (Central Processing Unit)
  • GPU - (Graphics Processing Unit - graphic processing unit or "graphics processing device")
  • now - APU (Accelerated Processing Unit - processor with video accelerator or "accelerated processor", if verbatim)

Purely technically, any processor with integrated graphics core can be referred to the APU. Another thing is that this abbreviation is not yet widely used, although it has already reached the point where the proportion of the graphic component in the video core reaches almost 50 percent of the area of ​​the common crystal.

Later, the FIR of Intel released its solutions with integrated graphics. Representatives are APU "Sandy Bridge" based platforms and their modification - "Ivy Bridge" for socket LGA 1155.

How is this possible: more and more components and functions are integrated into the same volume of the silicon crystal? Everything is simple - a constant reduction of the technological process! For example, when switching to a new process technology (22 nanometers), more transistors can be placed on the same area (the smaller the detail, the more they can be shoved into the final product).

Note: a nanometer (the old name is "millimicron") is one billionth of a meter! One of the most common units of measurement, found in semiconductor manufacturing technologies.

It got to the point that on some modern laptops the cooling system is not at all classical. I will explain my idea: instead of the usual layout of the three radiators on the heat pipes (processor, north and south bridge), we have one chip and one radiator! The heat sink tube runs through the entire board and ends with a fan that "ejects" heat outside the computer case .

On the motherboard, in fact, there are only various auxiliary elements: voltage drivers, systems for smoothing and filtering it (capacitors, chokes), elements responsible for starting the board and monitoring its temperature (multicontroller). All the main controllers and chips are assembled in one APU chip.

Here is the same mage-chip with the removed cooling system:

Obviously, this design gives minimal time delays in transferring data streams between all the nodes listed above. And this, in turn, provides very good performance for these integrated graphics cards. You can say this: the integrated graphics core in "Sandy Bridge" has the same performance as an external entry-level graphics card , or even more. And, of course, - hardware support for video stream in HD quality! :)

I want to add a small remark: with the traditional approach it is considered that the built-in graphics card can not work in any way with the external one, nor expand or supplement its functionality. Either that or another. If there is an external built-in graphics card simply turns off.

But, as they say, there are exceptions to any rule. In our case, these are laptops with two video cards. The first and the main, as a rule, there is some modification of "Intel HD". The second video card puts a more powerful graphics card from "AMD" or "NVidia". It is included in the work when the first (integrated) graphics card "does not pull". In three-dimensional games, for example.

You ask, why not put one powerful video card that will cope with all the tasks? And power consumption? It's a laptop, it's important for him to work as long as possible from the battery, and a powerful video card and energy consumes a lot. So manufacturers went for such a compromise. While you are working in Word or OpenOffice, an image on the screen displays an economical Intel graphics card. The toy was launched, - a powerful 3D-GPU from "AMD" or "NVidia" was switched on, having temporarily suspended the main one.

But here, again, the video cards work in turn, although they learned how to switch between themselves automatically and without rebooting. In fact, the joint work of the built-in and discrete video cards began with the advent of NVidia Optimus technology. In it, the integrated Intel graphics card is not turned off, but provides its own buffer buffer for a discrete neighbor. This is cooperation. A powerful card from NVidia forms an image and "puts it" in the frame buffer of the built-in card, and that takes responsibility for displaying the image on the screen.

So if you have a powerful discrete graphics card on your laptop, and games with 3D graphics are very slow, first check whether the laptop switches to it. Maybe the weak intel is trying to calculate 3D, but it turns out, all the same, not very much.