Accelerate the simulation with CUDA
Dear all,
I’m using esys to simulate some mechanical experiments such as Direct Shear and Uniaxial Compression of geotechnical materials. There are at least 100,000 particles in the model, as a result, when the number of timestep is large it takes several days or even more to complete one simulation although I have used the parallel calculation function with a desktop computer (8 cores ; Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz 3.70GHz) . So I’m considering rewrite the mainly time-consuming part of code in the calculation with CUDA to achieve a much higher efficiency, for example the part of contact detection. Now I have several questions to consult: 1、Whether this idea is feasible and what degree of workload will be? 2、If this is possible, base on the Ubuntu version or base on the Windows version which is more convenient? 3、If I want to speed up the part of contact detection may I only need to rewrite the corresponding code of that part? And which file is this part of code in?
Thanks a lot for your help in advance.
Ocean
Question information
- Language:
- English Edit question
- Status:
- Solved
- Assignee:
- No assignee Edit question
- Solved by:
- SteffenAbe
- Solved:
- Last query:
- Last reply: