Before we start off on our quest to gauge the CPU vs GPU it is good to get a rough understanding of what we are trying to compare or differentiate depending on how you look at it.
The Central Processing Unit, better known as the CPU has been with us for many decades. So much so that we have become ever so too familiar with its presence with us. CPUs can be found in almost every electronic device today, from computers to mobile and embedded devices.
The GPU, on the other hand, stands for Graphics Processing Unit. It is also lesser known as a visual processing unit (VPU). Basically, there are a number of similarities and assumptions between a CPU vs GPU. For starters, we can see that they are both processing units.
A Case to Compare the CPU vs GPU
Before we can start to dissect the two, we need to understand what each of them is. In the process of evaluating the two, we will automatically start to understand the similarities and differences between both of these electronic components.
For starters, a CPU can be described as an electronic circuitry that processes instructions. These instructions come from a program in the form a programming language such as machine language. The circuitry can be in several forms such as vacuum tube based transistors as in the olden days. Modern CPUs are made of electronics consisting of micro silicon-based transistors and wiring.
The CPU performs these instructions in the form of arithmetic, logic, control, input, and output. With that in itself, we can begin to suspect the CPU is not ideally optimized to perform certain tasks. In many computer systems additional processing in terms of audio synthesis and graphics acceleration or rendering required more dedicated hardware. These could be additional microprocessors controller units or entire dedicated hardware specialized devices may be required for these functions.
Taking Care of the Complexities
As roles became more demanding computers found a need to include expansion slots to aid in easing the burden of the CPU. These slots were the basis of extended cards handling Network I/O and dedicated cards to handle audio. One of the extended need came in the graphics arena and thus fueled the growth for GPUs.
GPUs so happen to be specialized CPUs that are designed to handle graphics capabilities. Modern GPUs even come with their own dedicated RAM and even have extended their operational role to complement the computers primary CPU.
GPUs which can be found in embedded systems, servers, mobile devices, personal computers, workstations, and game consoles are designed to render and optimize the output of graphic images whether or not in real-time.
Not only are GPUs restricted to graphics, but they can also be used to accelerate high-end computational tasks thereby allowing a computer to processes lots of computational data. These units are designed to be highly parallelized out of the box. They are designed to handle certain types of mathematical algorithms ideal to image processing and graphics rendering. They have the ability to buffer frames, matrix and vector calculations and all imaging-related stuff.
GPUs are not designed to work standalone unless in highly specialized circumstances. They are designed to complement the CPU and take a load off on the heavy computations tasks thus freeing the CPU for the more operating system, text processing, scheduling, and other tasks you would find on your computer.
Introduction to the GPU: http://www.nvidia.com/object/gpu.html
A brief history of computing