OpenMP and SIMD Instructions on Intel Xeon Phi

"Vector instruction sets have progressed over time, and it important to use the most appropriate vector instruction set when running on specific hardware. The OpenMP SIMD directive allows the developer to explicitly tell the compiler to vectorize a loop. In this case, human intervention will override the compilers sense of dependencies, but that is OK if the developer knows their application well."

The post OpenMP and SIMD Instructions on Intel Xeon Phi appeared first on insideHPC.