[QE-users] [QE-GPU] compilation with SDK 21.5 and CUDA 11.4

[Marcelo Falcão de Oliveira]

Dear QE users,

   I'm trying to compile pw.x with SDK 21.5 and CUDA toolkit 11.4 regarding
2 Tesla K20c gpus (Intel Xeon E5-2650 processors).
   I know the gpuss and packages are working because I was able to compile
lammps with them and it works. I have also compiled the test examples
provided by Nvidia for SDK and CUDA, including CUDA-Fortran, but I get the
following error when trying to configure. I'm probably making some bad QE
configuration but could not realise which one... Any hints? Thanks in
advance!

Configure command:
./configure --with-cuda=$CUDA_HOME --with-cuda-runtime=11.4
--with-cuda-cc=35 --enable-openmp --with-scalapack=no

Last output lines of configure:
using F90... nvfortran
setting FFLAGS... -O1
setting F90FLAGS... $(FFLAGS)
setting FFLAGS_NOOPT... -O0
setting CPP... cpp
setting CPPFLAGS... -P -traditional -Uvector
setting LD... mpif90
setting LDFLAGS...
checking for Fortran flag to compile .f90 files... (cached) none
checking whether Fortran compiler accepts -Mcuda=cuda11.4... no
configure: error: You do not have the cudafor module. Are you using a PGI
compiler?

-- 
Marcelo Falcão de Oliveira
Professor
Departamento de Engenharia de Materiais
EESC - USP
Av. João Dagnone, 1100,  Jd. Sta Angelina
CEP: 13563-120, São Carlos - SP - Brasil
e-mail: marcelo.falcao at usp.br
Fone: (16) 3373-9583
Fax: (16) 3373-9590
Agenda:
http://www.google.com/calendar/embed?src=marcelo.falcao%40usp.br&ctz=America/Sao_Paulo&mode=week
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210716/c01036fa/attachment.html>