CHAPTER 1 INTRODUCTION 1.1 Motivation The performance of the single core processor has hit the wall because of power requirements and heat dissipations. Then Hardware industry started creating multicore CPUs. Although, they can compute millions of instructions per second, there are some computational problems that are so complex that a powerful microprocessor needs years to solve them. To build more powerful microprocessors requires an expensive and intense production process. Some computations take years to solve even with the more powerful microprocessor. May be because of these factors, programmers sometimes use a different approach called parallel processing. Parallel processing defines that two or more microprocessors handle parts of …show more content…
Therefore resulting performance increases are generally less in magnitude. Even multicore processor lag behind the GPUs floating point operations. This is because CPUs has large portion of die space to control logic and cache memory. Whereas, GPU allocates large die space to arithmetic units .Thus GPUs are able to perform floating point operation faster. 1.2 Accelerated Computing on GPU A GPU is a specialized electronic circuit helps to compute graphical and non graphical computations faster. It acts as a co-processor to a conventional CPU to speed up the computations. A CPU takes more computation time then GPU for certain programs or tasks, which have large number of iterations. Because of large number of processor cores present in GPU. Modern GPUs are very effective at image processing and manipulating computer graphics, and their parallel architecture makes them more effective than general-purpose CPUs for algorithms where parallel processing of large blocks of data is required. In a personal computer, a GPU can be present on the motherboard , or it can be on a video card or on the CPU …show more content…
Not anymore, With CUDA, we can send C, C++ and FORTRAN code straight to GPU, no assembly language required. Using high-level languages, GPU-accelerated applications run the sequential part on the CPU – which is optimized for single threaded performance – while accelerating parallel processing on the GPU. This is called "GPU computing." 1.4 Objectives Convolution is one of the most important mathematical operations that are used in signal processing heavy application. In computer graphics and image processing fields, we usually work with discrete functions (e.g. an image) and apply a discrete form of the convolution to remove high frequency noise, sharpen details, detect edges, or otherwise modulate the frequency domain of the image. In this project, an efficient implementation of convolution filters on the GPU has been implemented. In this project convolution filter has been implemented in GPU using CUDA architecture and compared with the convolution implementation in CPU. CHAPTER 2 LITRATURE
Before we take a look at the CPU itself and its performance, it is worth noting that this CPU was manufactured at Globalfoundries, which is going to be AMD’s primary manufacturing facility for both CPUs and GPUs going forward. The new Thuban core may still be manufactured on the older 45nm node, however several improvements have been made to the manufacturing process and proof of this is how, despite having two additional cores, the total TDP has remain...
2. Cores are processors that combine two or more central processing units on one chip.
Process tuning strategy is to minimize the average power consumption of multi-core processors that use the DVFS and CG techniques, while providing the same maximum performance. By scaling the supply voltage (VDD), maximum operating frequency is increased to enhance the performance of a multi core processor. However, increasing VDD is limited by the power and thermal constraints. These constraints can be determined by capacity of supply voltage regulators and cooling solutions in the computing platforms customized for each computing segment.
Graphics card – The role of the graphics card is that it is a piece of hardware that will control the picture that will be outputted to the computer monitor, it will control the picture quality and the speed that certain graphics based programs will run at. This means that the graphics card is important when it comes to the computer as without it there would be no picture at all or the picture quality would be very poor and bad. The importance of the graphics card is also so when you are creating and editing graphical images that the graphical edits render in quicker and this will speed up the process of creating and editing graphical images considerably.
The Microsoft Xbox has powerful components within that overpower any other console, including its CPU or it’s Central Processing Unit. The Xbox runs an Intel Pentium III 733-megahertz (MHz) processor. This is a lot more powerful than Nintendo Gamecube’s IBM Power PC 485-megahertz Gekko chip processor and even lower the Sony Playstation 2’s 128 bit “Emotion Engine” which is clocked around 294-megahertz. A few more components that make this console far superior to any other is its graphics processor, or GPU, which stands for Graphic Processing Unit. Microsoft had teamed up with Nvidia to create a chip that was made especially for the Xbox. The computer-based company ‘s 250-megahertz chip gives the console an amazing resolution of 1920x1080 pixels and 125 Million Polygons per Second. Sixty four- (64) megabytes (Mb) of DDRam back all this up, for great non-glitch games and smooth running. Nintendo also has a computer-based company teaming up with them to create the 162-megahertz ATI Flipper GPU Chip. This chip can only handle around 6-12 Million polygons a second and is only backed up by 43-megabytes of DDRam.
The Cray X-MP/22 manufactured by Cray Research Incorporated (CRI) of Minneapolis, Minnesota was delivered and installed at the U of Toronto this September. The Cray is a well respected computer - mainly for its extremely fast rate of mathematical floating-point calculation. As the university states in its July/August computer magazine "ComputerNews", the Cray's "level of performance should enable researchers with large computational requirements at the university of Toronto and other Ontario universities to compete effectively against the best in the world in their respective fields." The Cray X-MP/22 has two Central Processing Units (CPUs) - the first '2' in the '22'. The Cray operates at a clock rate of 105 MHz (the regular, run-of-the-mill IBMPC has a clock rate of 4.77 MHz). By quick calculations, you would be led to believe the Cray is only about 20 times faster that the PC. Obviously, this is not the case.
The computer processor is the brain of the system. Our client requires a heavy punch, multi tasking brain. Due to his low income and student status, he does not have a large budget to put towards his computer. The AMD AthlonXP 1900 was chosen because of it's low cost and it's higher performance rating than that of the higher priced equivalent Pentium 4. The AMD AthlonXP 1900's speed is 1.6GHz. This means 1600Hz, which is a very fast processor. Amazingly, the clock speed of this processor is faster, albeit barely, than the Pentium 4 2.0GHz processor, and is much cheaper as well. This is important, as, for less cash, our student can afford a better processor.
The masks used for the convolution process is as shown below: The edge strength of the gradient in an image is then calculated by the formula : Step 3:- The gradient is the X and Y directions are used to calculate the direction of the edge. In the case of zero sum, X, an error will be generated. For cases where the sum equals zero, there has to be a restriction set.
Whenever you press a key, click the mouse, or start an application, you're sending instructions to the CPU. The CPU is generally a two-inch square with a silicon chip located inside. The CPU fits into the motherboard's CPU socket, which is covered by the heat sink, an object that absorbs heat from the CPU. A processor's speed is measured in megahertz (MHz), or millions of instructions per second; and gigahertz (GHz), or billions of instructions per second. A faster processor can execute instructions more quickly. However, the actual speed of the computer depends on the speed of many different components—not just the processor.
Most programming languages—such as C, C++, and Fortran—use compilers, but some—such as BASIC and LISP—use interpreters. An interpreter analyzes and executes each line of source code one-by-one. Interpreters produce initial results faster than compilers, but the source code must be re-interpreted with every use and interpreted languages are usually not as sophisticated as compiled languages.
Recently, a methodology for implementing lifting based DWT has been proposed because of lifting based DWT has many advantages over convolution based one [3-5]. The lifting structure largely reduces the number of multiplication and accumulation where filter bank architectures can take advantage of many low power constant multiplication algorithms. FPGA is used in general in these systems due to low cost and high computing speed with reprogrammable property.
In the past few decades, one field of engineering in particular has stood out in terms of development and commercialisation; and that is electronics and computation. In 1965, when Moore’s Law was first established (Gordon E. Moore, 1965: "Cramming more components onto integrated circuits"), it was stated that the number of transistors (an electronic component according to which the processing and memory capabilities of a microchip is measured) would double every 2 years. This prediction held true even when man ushered in the new millennium. We have gone from computers that could perform one calculation in one second to a super-computer (the one at Oak Ridge National Lab) that can perform 1 quadrillion (1015) mathematical calculations per second. Thus, it is only obvious that this field would also have s...
Processor speeds are measured in megahertz (MHz) and now come in speeds of up to 1000 MHz (1 GHz), which is very fast. This is almost ten times faster than the speed of most home computers, which average from 133 MHz to 166 MHz. Intel and AMD have been in a race to break the 1 GHz speed barrier, and the number of megahertz in the newest processors is not as significant as it was in earlier processors. For example, the difference between a 133 MHz processor and a 166 MHz processor is
So in order to give the CPU a break and help it run more efficiently, a Graphic card can be used to process the graphics portion of the processing load. Because most of today's programs are graphically oriented, the video card can help almost any program run more efficiently.
Software, such as programming languages and operating systems, makes the details of the hardware architecture invisible to the user. For example, computers that use the C programming language or a UNIX operating system may appear the same from the user's viewpoint, although they use different hardware architectures. When a computer carries out an instruction, it proceeds through five steps. First, the control unit retrieves the instruction from memory—for example, an instruction to add two numbers. Second, the control unit decodes the instructions into electronic signals that control the computer.