When using GPU computing, the number of CUDA cores is 0, and the GPU utilization is 1%-4%
I am using the version DualSPHysics_v5.2, the GPU version I am using is RTX 4070Ti, and CUDA 12.0 is also installed on Windows. But when I calculate, it shows CUDA cores as 0 and GPU utilization is very low, is this normal? Or is it that DualSPHysics_v5.2 does not support RTX 4070Ti graphics card computing?
To see the CUDA load, you have to change the graph via the dropdown in taskmanger.
Thank you for your reply, the problem has been solved, it turns out that because the current version does not support RTX 40 series graphics cards, I added a sentence to the code, recompile the solver, and now the utilization of the GPU can reach 100%.
Good you found the solution.
Sorry for late response... you can also send email to firstname.lastname@example.org if we do not reply here very fast.
@zhuzhu Hello, could you share the code which added to recompile the solver? I am using RTX 4090 for simulation and have the same question CUDA cores are 0 and GPU utilization is very low.