Topic
Parallel slowdown
About: Parallel slowdown is a research topic. Over the lifetime, 2 publications have been published within this topic receiving 10 citations.
Papers
TL;DR: A distributed-memory parallelization strategy for the density matrix renormalization group is proposed for cases where correlation functions are required and has substantial improvements with respect to previous works.
11 citations
1 Dec 2012
TL;DR: This work presents performance analysis of a parallel computing algorithm for deriving solar quiet daily (Sq) variations of geomagnetic field as a proxy of Space weather and found it was four times faster than the corresponding sequential algorithm under same platform and workload.
Abstract: This work presents performance analysis of a parallel computing algorithm for deriving solar quiet daily (Sq) variations of geomagnetic field as a proxy of Space weather. The parallel computing hardware platform used involved multi-core processing units. The parallel computing toolbox of MATLAB 2012a were used to develop our parallel algorithm that simulates Sq on eight Intel Intel Xeon E5410 2.33 GHz processors. A large dataset of geomagnetic variations from 64 observatories worldwide (obtained for the year 1996) was pre-processed, analyzed and corrected for non-cyclic variations leading to [366 x 276480] partitioned matrix, representing 101,191,680 measurements, corresponding to International Quiet Day (IQD) standard for the longitude, latitude and the local time of the individual stations. The parallel algorithm was four times faster than the corresponding sequential algorithm under same platform and workload. Consequently, Amdahl and Gustafson’s models on speedup performance metric were improved upon for a better optimal result.
1 citations
Related Topics (5)
Performance Metrics
| Year | Papers |
|---|---|
| 2012 | 1 |
| 2010 | 1 |