High-performance bidiagonal reduction using tile algorithms on homogeneous multicore architectures

Computer Science

• Multicore 100%
• Data Layout 100%
• Computational Science 33%
• Data Dependency 33%
• Memory Traffic 33%
• Main Memory 33%
• Bidiagonal Form 33%
• Total Execution Time 33%
• Data-Dependences 33%
• Open Source 33%
• runtime system 33%
• Processing Unit 33%
• Singular Value 33%
• Data Representation 33%
• Computer Software 33%

Keyphrases

• Homogeneous multicores 100%
• Two-stage Reduction 16%
• General Algorithm 16%
• Translation Layer 16%
• Reduction Approach 16%
• Total Execution Time 16%
• High Performance Kernels 16%
• Cache Sharing 16%
• Tile Size 16%
• Matrix Bandwidth 16%