A GPU is the chip that handles graphics and visual processing

A GPU (graphics processing unit) is a specialized chip inside your computer that draws everything you see on your screen. While your CPU (central processor) handles general computing tasks like running programs and managing files, the GPU focuses entirely on creating images, animations, and video. It processes millions of pixels simultaneously, which is why it exists as a separate piece of hardware rather than just part of the main processor.

Every computer has a GPU, though not all of them are obvious. Some are built directly into the CPU itself—these are called integrated graphics. Others are separate cards that plug into your motherboard—these are called discrete graphics cards or dedicated GPUs. A laptop or basic office computer usually has integrated graphics. A gaming computer, video editing workstation, or machine used for 3D design typically has a discrete GPU in addition to integrated graphics.

Key Takeaways

  • A GPU is a specialized processor that renders images, video, and animations on your screen by processing pixels in parallel.
  • Integrated GPUs are built into your CPU and share system memory; discrete GPUs are separate cards with their own memory and are much faster.
  • Gaming, video editing, 3D rendering, and machine learning all demand discrete GPUs because integrated graphics cannot process that volume of data quickly enough.
  • You can check which GPU your computer has by looking in your system settings or device manager, and you can upgrade a discrete GPU in a desktop but not in most laptops.

How a GPU differs from a CPU

Your CPU is designed to handle one task at a time very quickly. It excels at logic, decision-making, and running the operating system. A GPU, by contrast, is built to handle thousands of small identical tasks in parallel. Drawing a single frame on your screen requires calculating the color of millions of pixels—a GPU can do this by assigning each pixel to a different processing core and computing them all at once.

This difference in design makes GPUs terrible at general computing but exceptional at graphics. If you tried to use only a CPU to render a video game or edit 4K video, the computer would slow to a crawl. The GPU's parallel architecture is what makes these tasks feasible. Modern GPUs contain thousands of small cores, while CPUs typically have between 4 and 16 cores.

Integrated vs. discrete graphics

Integrated graphics are GPU cores built directly into your processor. They share your computer's main memory (RAM) and have no dedicated memory of their own. This saves space and power, which is why laptops and budget computers use them. Intel calls its integrated graphics "Intel UHD" or "Intel Iris." AMD calls theirs "Radeon." They handle everyday tasks like web browsing, document editing, and video playback without any problem.

Discrete graphics cards are separate components that plug into a slot on your motherboard. They have their own dedicated memory (called VRAM, usually 4 GB to 24 GB) and their own power supply. Because they do not share system RAM, they are much faster than integrated graphics. NVIDIA and AMD are the two main manufacturers of discrete GPUs. Discrete cards are necessary for gaming at high frame rates, professional video editing, 3D rendering, and machine learning work.

What tasks actually need a discrete GPU

Web browsing, email, spreadsheets, and document editing all run fine on integrated graphics. Streaming video and watching YouTube also work without a discrete GPU. The bottleneck appears when you try to render graphics in real time or process massive amounts of visual data.

Gaming is the most common reason people buy discrete GPUs. A modern game at high settings can demand 60 to 144 frames per second, with each frame containing millions of pixels. Integrated graphics cannot keep up. Video editors working with 4K or 8K footage face the same problem—the GPU must decompress, process, and re-encode video data faster than a CPU could manage alone. 3D designers using software like Blender or AutoCAD rely on discrete GPUs to render complex scenes. Machine learning and AI training also use GPUs because the math involved (matrix multiplication) maps perfectly onto parallel processing.

How to find out what GPU your computer has

On Windows, right-click your desktop and look for "NVIDIA Control Panel" or "AMD Radeon Settings." If neither appears, you have integrated graphics. You can also open Device Manager (search for it in the Start menu), expand "Display adapters," and look at what is listed. Integrated Intel graphics will say "Intel UHD" or "Intel Iris." A discrete card will show "NVIDIA GeForce" or "AMD Radeon" with a model number.

On Mac, click the Apple menu, select "About This Mac," and click "System Report." Under "Graphics/Displays" you will see your GPU listed. Most Macs use integrated graphics from Apple's own chips (M1, M2, M3) or older Intel integrated graphics. Some older Mac Pro models have discrete NVIDIA cards.

On Linux, open a terminal and type lspci | grep -i vga to see your graphics hardware. You can also type glxinfo | grep "OpenGL renderer" to see what the system is currently using for graphics.

Can you upgrade your GPU

In a desktop computer, upgrading a discrete GPU is straightforward. You remove the old card from its slot, install the new one, and read the latest drivers from NVIDIA or AMD. This takes about 15 minutes. You cannot upgrade integrated graphics because they are soldered to the CPU.

In a laptop, you almost never can upgrade the GPU. Laptop GPUs are either integrated into the CPU or soldered directly to the motherboard. Some older gaming laptops had removable discrete cards, but this is rare now. If you need a faster GPU and own a laptop, your only option is to buy a new computer.

In a Mac, GPU upgrades are not possible. The GPU is part of the chip itself and cannot be replaced without replacing the entire logic board.

GPU memory and performance

Discrete GPUs have their own memory, called VRAM (video RAM). Common amounts are 4 GB, 6 GB, 8 GB, 12 GB, and 24 GB. More VRAM allows the GPU to store larger textures, higher-resolution images, and more complex 3D models without having to swap data in and out of system RAM, which is much slower.

For gaming at 1080p resolution, 4 GB to 6 GB of VRAM is usually enough. For 4K gaming or professional work, 8 GB to 12 GB is more appropriate. Video editors and 3D artists working with very large projects sometimes need 24 GB or more. The GPU's clock speed (measured in MHz or GHz) also affects performance—higher clock speeds mean faster processing, but the architecture and number of cores matter more than raw clock speed alone.

Frequently Asked Questions

Do I need a discrete GPU if I only browse the web and watch videos?

No. Integrated graphics handle web browsing, streaming video, and document work without any problem. A discrete GPU would be a waste of money and power for these tasks. You only need one if you game, edit video, do 3D design, or work with machine learning.

What is the difference between NVIDIA and AMD GPUs?

Both make fast, reliable discrete graphics cards. NVIDIA has a larger market share and better support for gaming and machine learning. AMD cards are often less expensive for the same performance. For gaming, either brand works well. For machine learning, NVIDIA has more software support. Choose based on your budget and what software you plan to use.

Can I use my GPU for something other than graphics?

Yes. GPUs are used for machine learning, cryptocurrency mining, scientific simulations, and video encoding. Any task that involves processing large amounts of data in parallel can benefit from a GPU. This is why GPUs have become important for AI and data science work.

Will a better GPU make my computer faster at everyday tasks?

No. A GPU only speeds up graphics-related work. For web browsing, email, and office tasks, your CPU and RAM matter much more. A discrete GPU will not make these tasks noticeably faster and will only add cost and power consumption.

How hot do GPUs get, and is that a problem?

GPUs can run at 70 to 85 degrees Celsius under heavy load, which is normal. Most cards have built-in thermal throttling that reduces performance if they get too hot, protecting the hardware. As long as your case has adequate airflow and fans, a discrete GPU will not overheat during normal use. If your computer shuts down unexpectedly during gaming, overheating is a possible cause—check that your case fans are working.