GRAPHICS PROCESSING UNIT (GPU)

A graphics processing unit is a specialized circuit or a single processor designed to accelerate the image output in a frame buffer intended for output to a display .graphics processing units (GPUs) are very efficient of manipulating computer graphics and are generally more effective than general purpose CPUs at matrix manipulation. GPU algorithms allow processing of large blocks of data to be done in parallel.

GPU stands for general purpose computation on graphics processing units also known as GPU computing. PPUs are high performance, many core processors capable of very high computation and data through put. Once specially designed for computer graphics and difficult to program, todays GPUs are general purpose parallel processors with support for accessible programming interface and industry standard language such as C. developers who port their applications to GPUs often achieve speed up of orders of magnitude versus optimized  CPUs implementations. A GPU can contain thousands of cores, with cores inside cores, which perform a variety of small calculations concurrently.

A graphics processing unit (GPU) consists of several components including; cuda cores/stream processors, memory, texture mapping units (TMUs), and render outputs (ROPs).

The cuda cores/stream processors are the processing units that execute instructions and perform calculations.

The memory is the GPUs dedicated memory and it is used to store data and textures.

The texture mapping units (TMUs) handle texture mapping, which is the process of applying texture to 3d models.

The render outputs (ROPs) units handle the final rendering of pixels on the screen.

The GPUs process cycle involves the following; data transfer, instruction fetch, instruction decode, execution, texture mapping, rendering, memory access and data transfer.

The data transfer cycle starts with the CPU transferring data to the GPUs memory.

The instruction fetch cycle involves the GPU fetching instructions from the CPU.

The instruction decode cycle involves the GPU decoding the instructions and determining what operations is needed to be performed.

The execution cycle involves the cuda core/stream processors executing the instructions and performing calculations.

The texture mapping cycle involves the TMUs applying texture to the 3d models.

The rendering cycle involves the ROPs rendering the final pixels on the screen.

The memory access cycle involves the GPU accessing its own memory to retrieve data and textures.

Finally the data transfer cycle involves the GPU transferring the rendered pixels to the display device.

The advantages of GPUs are as follows; GPUs are designed to handle large amounts of data and perform complex calculations quickly. GPUs can perform many calculations simultaneously making them ideal for tasks that require parallel processing. GPUs can handle high resolution graphics and display them smoothly. Also GPUs are designed to be energy efficient which makes them suitable for use in laptops and other mobile devices.

The disadvantages of GPUs are as follows; high end GPUs can be very expensive making them inaccessible to some users. GPUs can generate a lot of heat which can lead to overheating and damage to the device. GPUs can consume a lot of power which can lead to high energy cost and environmental concern. GPUs can have compatibility issues with certain software and hardware which can lead to errors and crashes.

GPUs find application in a wide range of devices such as gaming consoles and workstations and professional CAD, graphics and animation workstation where they provide high quality graphics, smooth game play, and enhanced video capabilities. They are also used for high end scientific simulations and artificial intelligence workstations to provide excellent scientific simulations, machine learning and natural language processing capabilities and other applications where high speed processing and parallel processing is necessary or desirable.

The future of GPUs is expected to follow the trends and developments in the gaming, graphics, CAD, animation, artificial intelligence, virtual reality and quantum computing, where they will be required to provide advanced gaming experience, 3D modeling, higher speed and parallel processing capabilities, smoother game play and high quality graphics and above all to enhance the simulations of complex systems using quantum computing.

 

SOURCES:

• The GPU handbook: Real-time rendering and simulations by N.W Bowen.
• GPU Accelerated computing by R. Lopez.
• GPU Pro: Advanced rendering techniques by A.K Peters.
• High performance computing: Programming and applications by Victor Eijkhout.
• An introduction to general purpose GPU programming by J. Sanders and E. Kandrot.
• History of the GPU-steps to innovation by Jon Peddie.