CUDA Fortran compiler option, and (kind=16)

Accelerated Computing HPC Compilers nvc, nvc++ and nvfortran
1

Hello everyone:

Does CUDA Fortran permit kind=16, please?
I’m trying to use it, but keep getting

KIND parameter has unknown value for data type

Also, gfortran has the -fno-range-check for very large integers. Is there an equivalent for nvfortran, please?

Thanks,
Erin

2

Hi Erin,

No, sorry we don’t support quad precision natively. .Though you can use the QD library which ships with the compilers (under the “compilers/extras/qd/”). Documentation is included as well.

Hope this helps,
Mat