High performance matrix inversion based on LU factorization for multicore architectures

Computer Science

• Multicore 100%
• Numerical Library 100%
• Energy Efficient 50%
• Memory System 50%
• Power Consumption 50%
• Directed Acyclic Graph 50%
• Parallelism 50%
• Reported Result 50%
• Underlying Hardware 50%
• Algorithmic Approach 50%
• Synchronization Point 50%
• Inversion Procedure 50%
• Data Dependency 50%
• Computational Task 50%
• Runtime Environment 50%
• Data Layout 50%
• Computer Architecture 50%
• Data Programme 50%
• Processing Unit 50%
• Linear Systems 50%
• Efficient Implementation 50%

Keyphrases

• LU Factorization 100%
• Performance Matrix 100%
• Four-step 50%
• Utility Factor 25%
• ScaLAPACK 25%
• Program Data 25%
• Cache Performance 25%
• Power Consumption Analysis 25%
• Synchronization Point 25%
• Algorithmic Approach 25%
• Computational Tasks 25%
• Upper Triangular 25%
• Data Layout 25%
• Tile Data 25%
• Quark 25%
• Socket 25%
• System Memory 25%
• Runtime Environment 25%
• Loose Synchronization 25%
• Square Matrix 25%
• Column pivoting 25%