Abstract
The biggest characteristic of LogGP model based on LogP mode is sending long messages, if all the elements to be sent are seem as a long message and sent in a single processor, a sorting algorithm should be introduced to merge those elements, but the algorithm designed in the LogP mode is heavily dependent on the accuracy of parameters such as l, o, g, p. However, parameters are often inaccurate in reality. This may lead to message traffic congestion in the transfer process and the degradation of Communication performance of system. Therefore, this study proposes a new algorithm, that is, synchronization barrier is introduced into PSRS algorithm, which can improve LogGP Model further. Network congestion will be avoided when sending a long message and system performance will be improved .The barrier synchronization method can be applied to other algorithms of LogGP model, so it has a certain practicality.
Highlights
Many parallel computers provide special support for processing long messages, so a new model called LogGP was formed by introducing processing of long messages into LogP model
This study proposes a new algorithm, that is, synchronization barrier is introduced into PSRS algorithm, which can improve LogGP Model further
Network congestion will be avoided when sending a long message and system performance will be improved, the barrier synchronization method can be applied to other algorithms of LogGP model, so it has a certain practicality
Summary
Many parallel computers provide special support for processing long messages, so a new model called LogGP was formed by introducing processing of long messages into LogP model. This study proposes a new algorithm, that is, synchronization barrier is introduced into PSRS algorithm, which can improve LogGP Model further. Network congestion will be avoided when sending a long message and system performance will be improved, the barrier synchronization method can be applied to other algorithms of LogGP model, so it has a certain practicality. A synchronous technology based on the sharing storage model is often used in many parallel processing systems; it coordinates parallel computing by setting synchronous barricades. An application process can set many synchronous barricades points to coordinate parallel computing steps. The main idea of the barricade synchronization method is that all processes can continue to execute until they arrive at synchronous interaction point.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have