Cannot compile Fortran standard parallelism code for AMD GPU

[sumseq]

Hi,

I am trying the new amdflang compiler but cannot seem to get it to work.
I am able to use it to compile our code for the CPU, but I cannot activate "do concurrent" offload/threading.

Please see this issue I submitted to rocm llvm for details:

ROCm/llvm-project#214

The code is: github.com/predsci/hipft

• Ron

[jychang48]

Hi,

Thanks for reporting. The "do concurrent" feature not working is a known issue - the compiler team is working on it. I don't have an ETA of when this feature will work.