I am trying to build a Windows QE binary by cygwin64, OpenMP and gfortran ...
I was successful, but I see the performance is pure, result of my comparison tests bellow. Simple PBE L-alanine SCF calculation. The comparison is done with QE 6.4.1 build (made by others I am not able to replicate) ... I lose speed by a factor of 3x ... My effort to make an Intel Fortan + MKL Windows build of QE failed - the ./configure identify ifort, but rejects is as not suitable ... Any advice ? Anybody with Windows QE rebuild experiences available ? Michal Husak UCT Prague 1) QE 7.0, gfortan, 1 core + OpenMP, QE build in fft PWSCF : 30m15.81s CPU 30m58.45s WALL 2) QE 7.0, gfortan, 4 core + OpenMP, QE build in fft PWSCF : 34m20.25s CPU 21m25.60s WALL 3) QE 6.4.1, ifortran, 1 core + InptelMP, MKL PWSCF : 19m46.47s CPU 20m17.47s WALL 4) QE 6.4.1, ifortran, 4 core + InptelMP, MKL PWSCF : 37m 1.19s CPU 9m 7.10s WALL _______________________________________________ Quantum ESPRESSO is supported by MaX (www.max-centre.eu) users mailing list users@lists.quantum-espresso.org https://lists.quantum-espresso.org/mailman/listinfo/users
