English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the zaia usage as premium feature

ZAIA

Presents the pdf analysis as premium feature

PDF Analysis

Insights

Presents the summarisation as premium feature

Summarisation

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the access of premium content as premium feature

Premium Content

Zendy Tools

Zendy Open

This work describes the SIMD vectorization of the force calculation of theLennard-Jones potential with Intel AVX2 and AVX-512 instruction sets. Since theforce-calculation kernel of the molecular dynamics method involves indirectaccess to memory, the data layout is one of the most important factors invectorization. We find that the Array of Structures (AoS) with padding exhibitsbetter performance than Structure of Arrays (SoA) with appropriatevectorization and optimizations. In particular, AoS with 512-bit width exhibitsthe best performance among the architectures. While the difference inperformance between AoS and SoA is significant for the vectorization with AVX2,that with AVX-512 is minor. The effect of other optimization techniques, suchas software pipelining together with vectorization, is also discussed. Wepresent results for benchmarks on three CPU architectures: Intel Haswell (HSW),Knights Landing (KNL), and Skylake (SKL). The performance gains byvectorization are about 42\% on HSW compared with the code optimized withoutvectorization. On KNL, the hand-vectorized codes exhibit 34\% betterperformance than the codes vectorized automatically by the Intel compiler. OnSKL, the code vectorized with AVX2 exhibits slightly better performance thanthat with vectorized AVX-512.

SIMD Vectorization for the Lennard-Jones Potential with AVX2 and AVX-512 instructions