Accelerate LS-DYNA Software with the Latest Intel® Technologies
(LSTC) partnered to optimize LS-DYNA software, and it demonstrates the impact
of 2nd Generation Intel® Xeon® Scalable processor technologies for enhancing
the performance of LS-DYNA software. The paper focuses on two different
approaches for making use of Intel Advanced Vector Extensions 512
(Intel AVX-512) with LS-DYNA:
• LS-DYNA Explicit using Intel compiler vectorization techniques
• LS-DYNA Implicit using Intel Math Kernel Library (Intel MKL) for accelerating
dense matrix computational kernels
The paper also explores the numerical accuracy of simulation results for LS-DYNA
Explicit, comparing Intel Streaming SIMD Extensions 2 (Intel SSE2) and Intel AVX-
- Finally, the paper examines the benefits of Intel® Optane™ persistent memory
(PMem) for LS-DYNA Implicit simulations. The findings in this paper are drawn
from TopCrunch benchmarks (ODB-10M and car2car models) for LS-DYNA Explicit
and from AWE benchmarks (CYL1E6 and CL2E6 models) for LS-DYNA Implicit.
