Graphics Card

A graphics card is a fundamental hardware component in a computer system responsible for processing graphical data and transmitting it to display devices. Also known as a graphics processing unit (GPU), graphics cards play a central role in applications requiring high-resolution imagery. These units are essential for graphically intensive tasks such as 3D modeling, video editing, gaming technologies, and engineering applications.

Graphics cards are generally divided into two main categories: integrated (onboard) and discrete (dedicated) graphics cards. Integrated graphics cards are typically located on the same chip as the central processing unit (CPU) and share system memory. While sufficient for basic graphical operations, they offer limited performance. In contrast, discrete graphics cards have their own memory unit (VRAM) and are installed independently on the motherboard. Thanks to these features, they can achieve significantly higher graphics processing capacity.

Modern graphics cards are not limited to image generation alone; they also provide parallel processing power and are used in fields such as scientific computing, artificial intelligence applications, and cryptocurrency mining. The GPU architecture offers a significant advantage over CPUs in multi-data processing due to its structure with thousands of cores.

The key factors determining graphics card performance include the number of cores, clock speed, VRAM capacity, and memory bus width. Additionally, the cooling system and power consumption profile directly affect overall performance. Cooling systems are designed with either fan-based or liquid cooling solutions to maintain the card’s operating temperature within acceptable limits. In applications requiring high performance, advanced cooling solutions have become essential.

Integrated (Onboard) Graphics Card

An integrated GPU (English: Integrated GPU or IGP) is a graphics processing component embedded within the central processing unit (CPU) or on the chipset of the motherboard. This design performs basic graphical operations without requiring a separate graphics card and shares system memory (RAM) for processing graphical data.

Modern integrated GPUs have been developed by various manufacturers, including Intel’s HD/UHD Graphics, Iris, Iris Plus/Pro, and Xe series; AMD’s APU (Accelerated Processing Unit) architecture; and ARM-based Adreno and Mali GPUs. In particular, newer-generation integrated GPUs such as Intel Iris Xe offer significantly higher energy efficiency and graphical performance compared to previous models, providing an experience approaching that of a discrete GPU for low-to-mid-range gaming and video processing tasks.

Integrated GPUs do not contain dedicated VRAM since they rely on system RAM. In most systems, the amount of shared memory is dynamically determined by the BIOS or operating system. Thanks to this simple hardware architecture, integrated GPUs reduce the physical size of the system and ensure low power consumption. As a result, they are highly suitable for laptops, ultrabooks, mini-PCs, and embedded devices.

Advantages

The primary advantages of integrated GPUs include low cost, energy efficiency, and a compact form factor. Since no external graphics card or cooling system is required, device costs are significantly reduced. Additionally, low power consumption extends battery life and minimizes heat generation.

Furthermore, the use of shared system memory eliminates the need for additional physical space within the device, which is a significant advantage in portable systems. For everyday usage scenarios such as web browsers, office applications, and video playback, performance is adequate, and many users may not require a separate GPU.

Limits and Performance Constraints

The most important limitation of integrated GPUs is their reliance on shared system memory. Since system RAM is typically used concurrently for CPU operations, it offers lower bandwidth and higher latency compared to discrete GPUs. In modern systems, shared memory bandwidth may reach up to 128 GB/s, whereas VRAM data buses in discrete GPUs can provide bandwidth of 1000 GB/s or higher.

In such architectures, graphically intensive tasks such as high-resolution gaming, professional video editing, CAD, and 3D modeling suffer from insufficient performance. Integrated GPUs are limited in shader performance, thread count, and parallel core count, leading to bottlenecks under heavy graphics or data processing loads.

Use Cases

Integrated GPUs are preferred primarily in the following scenarios:

• Office software, email, and web browsing
• Video and media playback, with 1080p and most 4K content easily viewable
• Light photo editing and basic graphic applications
• Simple or older games at low to medium graphics settings
• Portable and slim devices, where energy efficiency is critical

Discrete (Dedicated) GPU

A discrete GPU (English: discrete GPU or dedicated GPU) refers to a separate graphics processing unit that operates independently of the CPU and is installed on the motherboard. Such GPUs feature their own dedicated memory (VRAM), power supply, and typically advanced cooling systems. This architecture enables them to deliver significantly higher graphics and parallel processing capacity than integrated GPUs.

Discrete GPUs are preferred in modern gaming computers, professional content creation systems (video editing, 3D modeling, CAD), artificial intelligence/machine learning applications, scientific computing centers, and high-resolution multi-monitor setups.

Advantages

• High Performance: Discrete GPUs feature thousands of parallel cores and perform shader operations, 3D graphics rendering, video encoding, and AI training tasks much faster than integrated solutions. This results in higher FPS, smoother graphical experiences, and faster render times.
• Dedicated VRAM: High-speed dedicated graphics memory (e.g., 8 GB or 24 GB GDDR6/HBM) reduces the load on system RAM and significantly alleviates data transfer bottlenecks.
• Advanced Technologies: Features such as ray tracing, AI-assisted graphics enhancements like DLSS/FSR, multi-monitor support, and VR are supported by discrete GPUs.
• Upgradability and Flexibility: In desktop systems, users can switch between different brands or models, replace cards, or set up multi-GPU configurations, enhancing system flexibility.

Disadvantages

• High Power Consumption and Heat: Discrete GPUs consume significantly more power and generate substantial heat compared to integrated solutions. Therefore, powerful cooling systems (fan-based or liquid-cooled) and adequate power supplies are required.
• Cost Impact: A new and powerful discrete graphics card substantially increases system cost. Additionally, energy consumption and cooling hardware add further expenses.
• Size and Portability: Although compact models are available for laptops, desktop GPUs generally require considerable physical space and are unsuitable for thin and lightweight mobile devices.

Technical Specifications and Operating Principle

Discrete GPUs are connected to the motherboard via PCI Express (PCIe) slots and operate independently from the system with their own processor unit, VRAM, power delivery, and cooling components. The graphics data processing pipeline consists of steps such as vertex processing, rasterization, fragment processing, and final image output. GPUs are optimized to execute these stages in parallel with high efficiency.

Through GPGPU usage, platforms such as CUDA or OpenCL enable parallel data processing for scientific simulations, machine learning training, and cryptocurrency mining at speeds far exceeding those of CPUs.

Use Cases

• Gaming and Entertainment: Delivers superior performance in AAA games at 1080p and higher resolutions with high FPS.
• Professional Content Creation: Reduces rendering times in software such as Adobe Premiere, Blender, AutoCAD, and Unreal Engine.
• Artificial Intelligence and Scientific Computing: Plays a critical role in fast processing of large datasets, AI model training, simulations, and engineering analyses.
• Multi-Monitor or 4K/8K Display: Provides stable performance and seamless multi-display support at high resolutions.

Comparison of Integrated (iGPU) and Discrete (dGPU)