Open Access. Powered by Scholars. Published by Universities.®
Physical Sciences and Mathematics Commons™
Open Access. Powered by Scholars. Published by Universities.®
- Discipline
- Institution
- Publication
Articles 1 - 4 of 4
Full-Text Articles in Physical Sciences and Mathematics
Understanding And Optimizing Parallel Performance In Multi-Tenant Cloud, Yong Zhao
Understanding And Optimizing Parallel Performance In Multi-Tenant Cloud, Yong Zhao
Computer Science and Engineering Dissertations
As a critical component of resource management in multicore systems, fair schedulers in hypervisors and operating systems (OSes) must follow a simple invariant: guarantee that the computing resources such as CPU cycles are fairly allocated to each vCPU or thread. As simple as it may seem, we found this invariant is broken when parallel programs with blocking synchronization are colocated with CPU intensive programs in hypervisors such as Xen, KVM and OSes such as Linux CFS. On the other hand, schedulers in virtualized environment usually reside in two different layers: one is in the hypervisor which aims to schedule vCPU …
Raspberry Pi Cluster For Parallel And Distributed Computing, Ngoc Ha
Raspberry Pi Cluster For Parallel And Distributed Computing, Ngoc Ha
Graduate Theses & Non-Theses
Parallel and distributed computing have become an essential part of the ‘Big Data’ processing and analysis, especially for geophysical applications. The main goal of this project was to build a 4-node distributed computing cluster system using the Raspberry Pi single-board computers for educational and research purposes. After assembling together the system, a standard test was performed to check the system functionality. A Monte Carlo simulation to calculate π (pi) was used to demonstrate the advantages and drawbacks of parallelization and distribution of tasks and data within the cluster. Challenges encountered during installation of the software and testing phase, and their …
A Study Of Several Applications Of Parallel Computing In The Sciences Using Petsc, Nicholas Stegmeier
A Study Of Several Applications Of Parallel Computing In The Sciences Using Petsc, Nicholas Stegmeier
Electronic Theses and Dissertations
The importance of computing in the natural sciences continues to grow as scientists strive to analyze complex phenomena. The dynamics of turbulence, astrophysics simulations, and climate change are just a few examples where computing is critical. These problems are computationally intractable on all computing platforms except supercomputers, necessitating the continued development of efficient algorithms and methodologies in parallel computing. This thesis investigates the use of parallel computing and mathematical modeling in the natural sciences through several applications, namely computational fluid dynamics for impinging jets in mechanical engineering, simulation of biofilms in an aqueous environment in mathematical biology, and the solution …
Scheduling Irregular Workloads On Gpus, David Arthur Troendle
Scheduling Irregular Workloads On Gpus, David Arthur Troendle
Electronic Theses and Dissertations
This doctoral research aims at understanding the nature of the overhead for data irregular GPU workloads, proposing a solution, and examining the consequences of the result. We propose a novel, retry-free GPU workload scheduler for irregular workloads. When used in a Breadth First Search (BFS) algorithm, the proposed simple, monolithic concurrent queue scales to within 10% of ideal scalability on AMD’s Fiji GPU with 14,336 active threads. The dissertation presents an important finding that the retry overhead associated with Compare and Swap (CAS) operations is the principle reason why concurrent queues do not scale well as the number of clients …