General-Purpose Computing on Graphics Processing Units (GPGPU) is an umbrella term for several technologies which offer several paradigms, directives, and primitives that play a huge role in accelerating compute-intensive applications by simply executing them on the GPU instead of the CPU. GPUs have thousands … Continue reading Implementing Breadth First Search in CUDA
CUDA is a parallel programming model developed by Nvidia to harness the power of the Nvidia-based GPUs so as to dramatically increase computing performance. Cuda is based on industry standard C / C++ as well as Fortran and provides a large set of primitives and … Continue reading Intro to CUDA
As part of a Parallel Computing course in my college, me and my classmates were supposed to always bring our thinking hats to the class and learn as well as solve various kinds of issues and problems associated with algorithms and implementation before they seem to … Continue reading Optimizing the Floyd Warshall Algorithm
I stumbled upon an already existing CALIC implementation on GitHub (btw. very rare to find such code) and decided to read through it. The author, Christopher A Taylor, attempted to implement it about 4 years ago, and it’s worth mentioning that he currently works in the startup GameClosure, which was famous for … Continue reading Parallelizing the implementation of CALIC
CALIC stands for Context-Based, Adaptive, Lossless Image Coding, and is an image codec that is made for obtaining a high degree of compression for continuous-tone images, and this compression is achieved with very low time and space complexities. CALIC puts heavy emphasis on image data modelling, and it … Continue reading An intro to the CALIC codec
Page Rank is a common term used to denote the famous algorithm that is incorporated by the Google Search engine to rank websites based on how important the website is. The metric here is basically the number of links to a web page that gives an … Continue reading Google Page Rank as a MapReduce Problem
If you’re someone who has worked with MPI (Message-Passing Interface) before, you would definitely agree on the fact that a cluster environment is necessary for learning MPI as well as developing and executing MPI-based applications, since a sole single machine does not always provide a … Continue reading Working with MPI in Docker
