Integrated graphics processor

from Wikipedia, the free encyclopedia
Intel 810e Northbridge with IGP

As an integrated graphics processor (short IGP , lehnübersetzt from the English integrated graphics processor ) are graphics processors called (GPUs) which is either:

  1. in the chipset of the motherboard - in this case a so-called onboard graphics card
  2. on a separate The - but in the same chip package as the main processor built in, or -
  3. directly on the die of the main processor and together with it (see also APU ).

The last new chipsets with onboard graphics cards - the AMD 800 series and the GeForce 320M - appeared in 2010. Integrated graphics processors have therefore been part of the main processor since then.

more details

A permanently installed or built-in ( integrated ) graphics processor normally uses the main memory of the computer because, in contrast to dedicated graphics cards , it usually does not have its own memory (more on the memory of dedicated graphics cards below, in the section on comparison with dedicated graphics cards ). The size of the so-called shared memory can usually be set in the computer's BIOS . Standard values ​​are 8  MiB to 1.5  GiB , although some systems can dynamically adjust the graphics memory used as required.

IGPs offer the same functions as dedicated graphics cards, but have a significantly lower speed for 3D presentations. With onboard GPUs, the connection with the Northbridge is implemented internally, as with dedicated graphics cards, via PCI , PCI Express or AGP ; Main processor-integrated GPUs are generally connected directly to the CPU.

They are often used in notebooks as well as office computers and servers.

An intermediate form to the dedicated graphics system are on-board graphics cards, which, however, have a separate graphics chip and their own memory, often found in Intel NUCs or gamer notebooks , for example ; there is occasionally also a combination of IGP (for economical everyday use) and additionally dedicated graphics (for 3D high-performance gaming) (see Nvidia Optimus , AMD Dynamic Switchable Graphics / Hybrid Graphics / "PowerExpress").

Comparison with dedicated graphics cards

In the following, the advantages and disadvantages between built-in graphics processors and dedicated graphics cards are compared. The latter usually consist of one or more GPUs which - in contrast to the integrated GPU - a private and dedicated account / dedicated ( dedicated ) or assigned work or graphics memory on their card have.

advantages

  • Does not require its own memory and hardly any printed circuit board space ;
  • Joint cooling with the CPU saves a separate cooling system;
  • Therefore, systems with IGPs are cheaper than with dedicated graphics hardware - if you also use the IGP.
  • Saving of space (for example with notebooks or tablets), as it is integrated in the northbridge of the chipset or in the processor;
  • Low power consumption - but correspondingly low performance;
  • Some IGPs have an additional L4 cache that can also be used by the CPUs.

disadvantage

  • A CPU with IGP and a supporting mainboard must exist for a desired computer system. This is often not available with workstation or server systems.
  • An IGP cannot be exchanged separately, often not at all. If the performance and connections are no longer sufficient, you can add a graphics card, provided there is a slot for graphics cards.
  • An IGP makes the CPU more expensive for users of separate graphics hardware.
  • Integrated GPUs usually do not have their own memory and therefore use the system memory. This memory is then no longer available to the system as working memory.
  • An IGP also uses part of the memory bandwidth of the CPU. This is then restricted for the CPU.
  • The IGP uses part of the max. Power consumption of the overall chip; the CPU may have to throttle so as not to exceed the joint thermal budget.
  • The 3D graphics performance of integrated GPUs is usually in the low-performance range, but is sufficient for desktop compositing , games with low to medium graphics requirements and video decoding.

Manufacturer

Well-known manufacturers of IGPs were AMD (formerly ATI Technologies ), Intel , Nvidia , SiS and VIA Technologies .

Widely used chipsets with onboard graphics cards were Intel Graphics Media Accelerator , Nvidia nForce and ATI Radeon Xpress . They were inherited by CPUs with Intel HD Graphics (including Iris, Iris Pro and UHD Graphics) and AMD Fusion . The successor to AMD Fusion are the Ryzen APUs.

Due to the widespread use of integrated graphics solutions, Intel is the market leader for graphics chips without having dedicated graphics cards on offer. In the third quarter of 2008, 49.4% (2006: 34.8%; 2003: 35%) of the graphics units sold in the overall market were attributable to Intel, 27.8% to Nvidia and 20.6% to AMD. In the fourth quarter of 2015, it was 71.6% for Intel, 16.6% for Nvidia and 11.8% for AMD.

See also

  • Embedded system - more general, not just limited to graphics [processor (s)]

Single receipts

  1. Borrowing from the English dedicated , which was probably already borrowed there in a similar way from the Latin dedicare .
  2. dedicated ( Memento of the original from August 16, 2016 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. (English-German) - Duden , Langenscheidt ; 2015 @1@ 2Template: Webachiv / IABot / www.duden.de
  3. dedicate - Duden , Bibliographisches Institut ; 2016
  4. Dedicated graphics memory - what is it? - Chip.de , on November 20, 2015.
  5. GPU MARKET DEFIES GRAVITY - so far ( memento of the original from October 30, 2008 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. (English) - Jon Peddie Research , October 27, 2008. @1@ 2Template: Webachiv / IABot / jonpeddie.com
  6. For the 4th quarter of 2015, GPU shipments increased 2.4% from last quarter ( Memento of the original from May 8, 2016 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. (English) - Jon Peddie Research ; 2016 @1@ 2Template: Webachiv / IABot / jonpeddie.com