Buzz
Graphics cards receive data from your computer's central processor (CPU) and convert it into visual output that appears on your screen. Image courtesy of SasinT Gallery/Getty ImagesThe images displayed on your computer screen consist of small dots called pixels. On most common resolution settings, your display shows over 2 million pixels, and the computer must process each one to form an image. To do this, it requires a translator—a device that converts binary data from the CPU into something you can see. This translator is called a graphics processing unit (GPU).
Many entry-level consumer laptops and desktops are equipped with an integrated GPU built into the main processor. However, professional-grade or custom systems often include space for a dedicated graphics card. The advantage of a dedicated graphics card is that it can render intricate visuals much faster than an integrated chip.
While the job of a graphics card is intricate, understanding its fundamental principles and components is straightforward. In this article, we'll explore the essential parts of a video card and their functions, as well as the key factors that contribute to a fast, efficient graphics card.
Imagine a computer as a company with its own art department. When employees need artwork, they submit a request to the art team. The department then determines how to create the image and produces it. The result is transforming an idea into a visible, tangible picture.
A graphics card follows a similar process. The CPU, working in sync with software, sends image-related data to the graphics card. The card then determines how to arrange the pixels on the screen to form the image and transmits this data to the monitor via a cable.
Generating an image from binary data is a challenging task. For a 3-D image, the graphics card starts by constructing a wireframe from straight lines. It then rasterizes the image (fills in the remaining pixels), adding lighting, textures, and colors. For fast-paced games, this process needs to occur 60 to 120 times per second. Without a graphics card to handle these calculations, the computer couldn't manage the workload.
The graphics card achieves this feat using four primary components:
- A connection to the motherboard for both data transfer and power supply
- A graphics processor (GPU) to determine how each pixel on the screen is handled
- Video memory (VRAM) to store pixel data and temporarily save completed images
- A monitor connection to display the final image
Next, we will dive deeper into the processor and memory components.
The Graphics Processing Unit (GPU)
A GPU is an electronic circuit in your computer that accelerates the creation and rendering of computer graphics. Image courtesy of CHALERMPHON SRISANG/ShutterstockSimilar to a motherboard, a graphics card is a printed circuit board that contains a processor and VRAM. It also includes an input/output system (BIOS) chip that holds the card’s settings and runs diagnostics on the memory, input, and output during startup.
The processor of a graphics card, called the graphics processing unit (GPU), is like a computer’s CPU. However, a GPU is specifically engineered for handling the complex mathematical and geometric computations needed for graphics rendering. Some of the most advanced GPUs have more transistors than typical CPUs.
A GPU generates a lot of heat, so it is typically placed beneath a heat sink or fan. Integrated chips, however, differ in that they lack dedicated VRAM and instead share the same RAM as the CPU. This may lead to memory shortages, especially during gaming, when using an integrated GPU.
Along with its processing power, a GPU uses specialized programming to analyze and manage data. AMD and Nvidia dominate the GPU market, with both companies developing unique features to enhance GPU performance. Today’s video processors can offer:
- Full scene anti-aliasing (FSAA), which smooths the edges of 3-D objects
- Anisotropic filtering (AF), which sharpens image quality
- Real-time physics and particle effects
- Support for multiscreen displays
- High frame rate video output
- Ultra high-definition video with millions of pixels
- GPU-accelerated calculations
Each company has also created unique methods to help the GPU apply colors, shading, textures, and patterns.
As the GPU generates images, it requires a place to store information and completed pictures. It utilizes the card's RAM for this task, holding data about each pixel, its color, and its position on the screen. Some of the VRAM also serves as a frame buffer, storing finished images until they are ready to be displayed. Video RAM typically operates at very high speeds and is dual-port, allowing the system to both read and write simultaneously.
Modern video cards are inserted into a PCIe x16 expansion slot. However, small form-factor devices with integrated graphics, such as laptops and mini desktops, might not have this slot. Graphics cards can still be connected through an expensive workaround known as an external GPU.
Graphics cards have evolved significantly since IBM introduced the first one in 1981. Known as the Monochrome Display Adapter (MDA), it displayed only text in green or white on a black screen. Today, both graphics cards and integrated chips can easily output HD (1,920 x 1,080 pixel) signals through HDMI or DisplayPort cables. Standalone cards often support Ultra HD 4K (3,840 x 2,160) video, with even higher resolutions available on more advanced GPUs.
How to Choose the Right Graphics Card
A top-tier graphics card is easy to recognize. It boasts ample memory and a speedy processor. It’s often more aesthetically striking than other components meant to go inside a computer’s case. Many high-performance video cards feature custom illustrations, or include decorative fans or heat sinks.
However, a high-end card offers more power than most users actually require. Those who use their computers mainly for email, word processing, or social media can rely on integrated graphics from the CPU to meet their needs. A mid-range card is usually sufficient for casual gamers. People who demand the power of a top-tier card include gaming enthusiasts and professionals working with 3-D graphics.
A key indicator of a card’s performance is its frame rate, expressed in frames per second (FPS). This metric reveals how many complete images the card can render each second. While the human eye can process around 25 frames per second, fast-paced games require at least 60 FPS for smooth animation and scrolling. The components that affect the frame rate include:
- Triangles or vertices per second: 3-D images are made up of triangles or polygons. This metric measures how fast the GPU can calculate the entire polygon or the vertices defining it. Essentially, it indicates how quickly the card constructs a wireframe image.
- Pixel fill rate: This metric shows how many pixels the GPU can process per second, essentially determining how quickly it can fill in the image with pixels.
The speed of a graphics card is directly influenced by its hardware. Here are the key hardware specifications that impact a card’s speed, along with the units in which they are measured:
- GPU clock speed (MHz)
- Memory bus width (bits)
- Available memory size (MB)
- Memory clock speed (MHz)
- Memory bandwidth (GB/s)
The computer’s CPU and motherboard also play a role in performance. Even a top-tier graphics card won’t reach its potential if the motherboard can’t transfer data quickly enough. Likewise, the connection between the card and motherboard, along with the speed at which it receives instructions from the CPU, significantly influences its efficiency.
For further details about graphics cards and related subjects, check out the following resources.
Many modern CPUs come with built-in graphics capabilities, allowing them to function without an additional graphics card. These processors are excellent for handling 2-D tasks and are perfect for productivity and web browsing. Installing a discrete graphics card will override the integrated graphics functions.
Some users choose to enhance their card’s performance by manually increasing the clock speed, a process called overclocking. Typically, overclocking is done to the memory, as increasing the GPU clock speed may result in overheating. While overclocking can improve performance, it also voids the warranty from the manufacturer.
