What are Dedicated & Integrated graphics cards?

Dedicated graphics cards

The GPUs of the very most powerful class typically interface with the motherboard by way of an expansion slot such as for example PCI Express (PCIe) or Accelerated Graphics Port (AGP) and may usually be replaced or upgraded with relative ease, assuming the motherboard is with the capacity of supporting the upgrade. A few graphics cards still use Peripheral Component Interconnect (PCI) slots, but their bandwidth is indeed limited they are generally used only once a PCIe or AGP slot isn't available.

Source: gpuhub render thue

A dedicated GPU isn't necessarily removable, nor does it necessarily interface with the motherboard in a typical fashion. The word "dedicated" refers to the actual fact that dedicated graphics cards have RAM that's focused on the card's use, never to the fact that a lot of dedicated GPUs are removable. Further, this RAM is normally specially selected for the expected serial workload of the graphics card (see GDDR). Sometimes, systems with dedicated, discrete GPUs were called "DIS" systems, instead of "UMA" systems (see following section). Dedicated GPUs for portable computers are mostly interfaced through a nonstandard and frequently proprietary slot because of size and weight constraints. Such ports may be considered PCIe or AGP when it comes to their logical host interface, even if they're not physically interchangeable with their counterparts.

Technologies such as for example SLI and NVLink by Nvidia and CrossFire by AMD allow multiple GPUs to draw images simultaneously for an individual screen, increasing the processing power designed for graphics. These technologies, however, are increasingly uncommon, because so many games usually do not fully utilize multiple GPUs, because so many users cannot afford them. Multiple GPUs remain applied to supercomputers (like in Summit), on workstations to accelerate video (processing multiple videos simultaneously) and 3D rendering, for VFX and for simulations, and in AI to expedite training, as may be the case with Nvidia's lineup of DGX workstations and servers and Tesla GPUs and Intel's upcoming Ponte Vecchio GPUs.

Integrated graphics processing unit

Integrated graphics processing unit (IGPU), Integrated graphics, shared graphics solutions, integrated graphics processors (IGP) or unified memory architecture (UMA) start using a part of a computer's system RAM instead of dedicated graphics memory. IGPs could be integrated onto the motherboard within the chipset, or on a single die with the CPU (like AMD APU or Intel HD Graphics). On certain motherboards, AMD's IGPs may use dedicated sideport memory. That is another fixed block of powerful memory that's dedicated for use by the GPU. In early 2007, computers with integrated graphics take into account about 90% of most PC shipments. They are less costly to implement than dedicated graphics processing, but tend to be less capable. Historically, integrated processing was considered unfit to play 3D games or run graphically intensive programs but could run less intensive programs such as Adobe Flash. Examples of such IGPs would be offerings from SiS and VIA circa 2004. However, modern integrated graphics processors such as AMD Accelerated Processing Unit and Intel HD Graphics are more than capable of handling 2D graphics or low stress 3D graphics.

Because the GPU computations are really memory-intensive, integrated processing could find itself competing with the CPU for the relatively slow system RAM, since it has minimal or no dedicated video memory. IGPs can possess up to 29.856 GB/s of memory bandwidth from system RAM, whereas a graphics card may have up to 264 GB/s of bandwidth between its RAM and GPU core. This memory bus bandwidth can limit the performance of the GPU, though multi-channel memory can mitigate this deficiency. Older integrated graphics chipsets lacked hardware transform and lighting, but newer ones include it.