I do hope it will be valuable for the community. Right now we are discussing internally which model to implement in the first place: k-omegaSST or k-epsilon. What is your opinion on the applicability on these models? BTW, could you fill in this survey? It should not take more than 3.44mins: http://goo.gl/LECn5o
Could you please share your experience by OpenFOAM and GPU acceleration?
I did some search and it seems to me that OpenFOAM has limitations about GPU programming. There are some libraries which they are not efficient enough and have some disadvantage such as CPU-GPU parallel programming because of different data structure and so on.
What do you think about GPU acceleration in OpenFOAM?
Do you think that we can be using OpenACC for GPU acceleration in OpenFOAM and getting good efficiency?