GPU-acceleration

[lsuser888]

Hi 
According to my research, GPU acceleration is not available for the explicit LS-DYNA solver. However, I came across the following article on the NVIDIA website:
https://developer.nvidia.com/blog/advancing-ansys-workloads-with-nvidia-grace-and-nvidia-grace-hopper/ 

The car crashing cannot be implicit so could you please clarify whether GPU acceleration is available for the explicit solver or not?

[peteroznewman]

Reading the article, the LS-Dyna solver was running on the Grace CPU, not the GPU.

[nguyenthuu30031999]

GPU acceleration in ANSYS can be quite complex, as support depends on the specific solver, license, and compatible hardware. It's helpful to see real user experiences and recommendations when optimizing simulation performance.

[Ram Gopisetti]

Hi

Based on the NVIDIA blog post, they ran Ansys LS-DYNA crash simulations on Grace CPU Superchip deliver over 2x performance per watt improvement compared to x86 alternatives, enabling up to twice the performance within the same power envelope not on GPU and using Multinode scaling tests show Grace CPU Superchip maintains strong scaling efficiency from 1 to 8 nodes for both ODM_10m and car2car_20m crash models against Intel Xeon Platinum 8480+.

So there is no GPU used here and currently we LS-Dyna dont support any GPU compatiblity for Implict or Explicit runs but we have some ongoing developemnt on this front. 

Cheers Ram