chrono
chrono copied to clipboard
How use CUDA acceleration in Chrono
First of all, I'am new in Chrono. Build Chrono 5 on Ubuntu 18 with flag to use CUDA and parallel module. All project compile without errors, found CUDA and other lib. But when run demos or compile individual samples from tutorial, ChSystemParallel use as understand omp system, it runs on 16 threads of CPU and don't use CUDA at all, on massive amont of bodies in example "cratering solution" from git. What flags need to use or what paramets and on what stage:
- In cmakelist of individual program
- In code of program
- When build Chrono ?