Acceleration of Apriori Algorithm on GPU using CUDA
Project Abstract / Summary : Data mining has become a hot research domain in recent years as it is being used in almost all applications. There are many algorithms under this domain. However these algorithms are potentially unable to handle today's increasing data sets or would take a tremendous amount of time. Therefore, users have to turn to rely on parallel and distributed computing techniques to accelerate the computation.
One of the parallel computing techniques is the General Process Computing on Graphics Processing Unit (GPGPU). This technique uses the power of GPU and CPU to perform computations on applications traditionally handled only by CPU.
A graphics processing unit (GPU) is a specialized electronic circuit designed to rapidly manipulate and alter memory to accelerate the creation of images in a frame buffer intended for output to a display. It allows blocks of data to be processed in parallel. Currently, high level languages have emerged to support easy programming on GPUs. CUDA is a parallel computing platform and programming model invented by NVIDIA. It enables dramatic increases in computing performance by harnessing the power of the graphics processing unit (GPU). Low cost is another factor of GPU. Therefore, recently, there has been a trend to accelerate computational intensive applications on a GPU + CPU heterogeneous system where the GPU acts as the computation accelerator. We propose a parallel-Apriori association rule mining algorithm using CUDA based technique. Comparison of the execution time of Parallel-Apriori with an efficient serial Apriori program is used to show the speedup on a real-world data set and even on synthetic data sets.
Why did you choose to work on this project topic : I found this project interesting because it optimizes an algorithm by considering minute details of the hardware alongside software. So there is a great scope to learn in depth about what is exactly happening and how can we minimize computing time of any algorithm. This concept can be applied in other fields like cryptography, image processing etc.
Project Category : CS / IT / Networking
------------------------------------------------------
Institute/College Name: Sreenidhi Institute of Science and Technology
City: Hyderabad
State: Telangana
Participating Team From: Final Year
One of the parallel computing techniques is the General Process Computing on Graphics Processing Unit (GPGPU). This technique uses the power of GPU and CPU to perform computations on applications traditionally handled only by CPU.
A graphics processing unit (GPU) is a specialized electronic circuit designed to rapidly manipulate and alter memory to accelerate the creation of images in a frame buffer intended for output to a display. It allows blocks of data to be processed in parallel. Currently, high level languages have emerged to support easy programming on GPUs. CUDA is a parallel computing platform and programming model invented by NVIDIA. It enables dramatic increases in computing performance by harnessing the power of the graphics processing unit (GPU). Low cost is another factor of GPU. Therefore, recently, there has been a trend to accelerate computational intensive applications on a GPU + CPU heterogeneous system where the GPU acts as the computation accelerator. We propose a parallel-Apriori association rule mining algorithm using CUDA based technique. Comparison of the execution time of Parallel-Apriori with an efficient serial Apriori program is used to show the speedup on a real-world data set and even on synthetic data sets.
Why did you choose to work on this project topic : I found this project interesting because it optimizes an algorithm by considering minute details of the hardware alongside software. So there is a great scope to learn in depth about what is exactly happening and how can we minimize computing time of any algorithm. This concept can be applied in other fields like cryptography, image processing etc.
Project Category : CS / IT / Networking
------------------------------------------------------
Institute/College Name: Sreenidhi Institute of Science and Technology
City: Hyderabad
State: Telangana
Participating Team From: Final Year
Replies
You are reading an archived discussion.
Related Posts
Project Abstract / Summary : MODIFICATIONS FOR THE IMPROVEMENT OF THE PERFORMANCE OF VORTEX TUBE
Vortex tube is a simple energy separating device which is compact and simple to produce...
Last year, NASA had announced its proposal to develop a conceptual design of a possible submersible autonomous vehicle (submarine) for performing autonomous detailed scientific investigations of the liquid methane/ethane seas...
Project Abstract / Summary : Our plan is to make a fighter jet which would have 2 special uses. They are:-
• Fast means of transport for military purpose.
•...
Digital Locker is one of the key initiatives under the Digital India Programme. A beta version (https://digitallocker.gov.in) of the same has been released on 10th Feb. 2015, by Department of...
Project Abstract / Summary : concrete is second most consumed materials in the world after water.concrete made up of cement,water ,fine and coarse aggregates.In order to make high strength concrete...