If you are on your path of learning about graphics of your computer, you certainly should know more about integrated GPUs and video memory and these are the things we are going to discuss in this article.
Video memory is one of the most crucial component of a computer system. This component is perhaps the most hard-working one in the system. I know you might start comparing with various other components as well. But all of them generally take rest in between and mostly works when requires. For instance, you grabbed your popcorn and beverage and about to sit for a TV series binge for next unknown hours. You need to double click the .MP4 or other video files and you are good to go.
Upon double clicking the file, you send information the CPU, RAM and HDD and they carry the task accordingly. Now check this out, for the video card, it has to constantly updates and reflect every changes that is taking place in the TV series and showing it on the display. Now, when you play heavy applications such as games, video memory can come out as a heavy bottleneck in three ways. Data throughput, access speed and simple capacity.
Manufacturers were able to overcome these bottlenecks by adding specialised super fast RAM and adding more and more total RAM. As a very first step, they reorganized the display memory on cards to use a wider bus, giving them more memory bandwidth. The system bus is limited to 32 or 64 bits and this would not come in with much of benefits if video display cards were not a coprocessor boards. All of or most of the graphics rendering and processing is handled by the video processor chip rather than by the CPU. The main system provides the gather input data straight to the processor on the video card. Since the memory bus on the video card can be wider many times than the standard 64-bit pathway, all these data can be manipulated and sent to the monitor much quickly.
Perhaps the last single most revolutionary invention in the computer system is the specialized dedicated video RAM. RAM offered radical improvements over its predecessors. One of the major distinctions between DRAM and VRAM is that video RAMs are able to read and write data simultaneously. There are various types of RAM you will find in the market. Like, DDR3 or recent generation of DDR RAM, GDDR3/4/5, GDDR5X, GDDR6, HBM, HBM2 etc.
There are motherboards which have integrated GPUs, and also quite a few are able to support CPU with an integrated GPU. The GPU on the motherboard can be a complete separated chip or could be built into the Northbridge chip. A GPU built into the motherboard is typically called an on-board video card. On-board video is not thus powerful but good for low demanding tasks and most common in laptops because it saves power and space.
Both AMD and NVIDIA make integrated GPU chips and Intel has already went down the road to integrated GPU called Intel Graphics Media Accelerator (GMA). There are fusion processor of AMD which have a CPU/GPU combination that stand tough against any CPU/Graphics card combination. That is typically at the level portable computers. They are good enough for casual gaming and a little medium duty gaming. As for Intel’s GMA, they are tailored for desktop performance and typical administrative task, not for gaming at all. There is NVIDIA’s Tegra, which is built for heavy duty gaming, also used in Nintendo Switch and automotive entertainment systems.
With the new concept of integrated GPU, the CPU section can get quite crowded. For instance, a single AMD Accelerated Processing Unit ( APU ) chip integrates around two to four CPU cores, a memory controller which supports DDR4 for the system memory, cache memory and a GPU that is able to handle advanced 3-D graphics. But all these integrated chips require way less electricity than discrete components.