Graphics processing unit

What is a Graphics Processing Unit (GPU)?

A graphics processing unit (GPU) is a specialized electronic circuit designed for digital image processing and computer graphics. Initially, GPUs were used for graphic calculations but later found to be useful for non-graphic calculations due to their parallel structure.

From Arcade Boards to Modern Marvels

Have you ever wondered how those early arcade games managed such impressive visuals? Since the 1970s, arcade system boards have used specialized graphics circuits. The first implementation of a personal computer graphics display processor was the NEC μPD7220, which enabled low-cost high-performance video graphics cards. Other notable developments include custom blitter chips in arcade games and the Amiga with its custom graphics chip.

Evolution Through the Decades

In the 1980s, various companies released innovative graphics processors such as the ARTC HD63484 and TMS34010. IBM released its proprietary Video Graphics Array (VGA) display standard in 1987, while Sharp’s X68000 used a custom graphics chipset with a 65,536 color palette.

The Rise of Dedicated GPUs

As we moved into the 1990s, dedicated polygonal 3D graphics boards were introduced in arcades. The first was the Namco System 21 and Taito Air System. IBM released its Video Graphics Array (VGA) display standard while NEC announced the creation of the Video Electronics Standards Association (VESA).

From S3 to NVIDIA

In 1991, S3 Graphics introduced the S3 86C911, which spawned a variety of imitators. The first hardware T&L GPU on home video game consoles was the Nintendo 64’s Reality Coprocessor in 1996. NVIDIA’s GeForce 256 was the first consumer GPU with hardware-accelerated T&L.

Programmable Shaders and Beyond

In the early 2000s, the GeForce 3 introduced programmable shading, allowing each pixel and geometric vertex to be processed by short programs. Early shader execution engines were not general-purpose but rather highly customizable function blocks.

Modern GPU Applications

The term ‘GPU’ was coined by Sony in reference to the 32-bit Sony GPU (designed by Toshiba) in the PlayStation video game console. In the PC world, notable failed attempts for low-cost 3D graphics chips included the S3 ViRGE, ATI Rage, and Matrox Mystique.

From Gaming to General Computing

The Nvidia GeForce 256 was the first consumer GPU to support programmable shading, used in Xbox and competing with PlayStation 2’s custom vector unit for vertex processing. With the introduction of the GeForce 8 series and generic stream processing units, GPUs became more generalized computing devices.

From Ray Tracing to AI

Nvidia’s CUDA platform was introduced in 2007, followed by OpenCL, an open standard that allows for development of code for both GPUs and CPUs with emphasis on portability. In the 2010s, Nvidia partnered with Audi to power cars’ dashboards using Tegra GPU.

Intel’s Re-entry

In 2022, Intel re-entered the discrete GPU market with its Arc series competing with GeForce 30 and Radeon 6000 series cards at competitive prices. GPUs are increasingly used for calculations involving embarrassingly parallel problems like training of large language models.

Performance Metrics

Performance of GPUs is affected by factors such as connector pathways, clock signal frequency, memory caches, streaming multiprocessors, compute units, and Xe cores. GPU performance is typically measured in floating point operations per second (FLOPS) or teraflops (TFLOPS).

Dedicated vs Integrated

Most GPUs are designed for specific uses such as real-time 3D graphics, gaming, cloud gaming, workstation, artificial intelligence training, automated cars, and more. Dedicated graphics processing units use RAM that is dedicated to the GPU, rather than relying on the computer’s main system memory.

External GPUs

An external GPU (eGPU) is a graphics processor located outside of the housing of the computer, sometimes used with laptop computers to enhance their graphics capabilities. Those ports are only available on certain notebook systems. eGPU enclosures include their own power supply (PSU), because powerful GPUs can consume hundreds of watts.

Conclusion

The journey from simple arcade boards to the complex, versatile GPUs we have today is nothing short of remarkable. From gaming to general computing and beyond, GPUs continue to evolve, pushing the boundaries of what’s possible in digital image processing and beyond. As technology advances, so too will the role of these powerful processors in our daily lives.