depot/third_party/nixpkgs/pkgs/development/libraries/elpa/default.nix

{ lib, stdenv, fetchurl, autoreconfHook, gfortran, perl
, mpi, blas, lapack, scalapack, openssh
# CPU optimizations
, avxSupport ? stdenv.hostPlatform.avxSupport
, avx2Support ? stdenv.hostPlatform.avx2Support
, avx512Support ? stdenv.hostPlatform.avx512Support
# Enable NIVIA GPU support
# Note, that this needs to be built on a system with a GPU
# present for the tests to succeed.
, enableCuda ? false
# type of GPU architecture
, nvidiaArch ? "sm_60"
, cudatoolkit
} :

assert blas.isILP64 == lapack.isILP64;
assert blas.isILP64 == scalapack.isILP64;

stdenv.mkDerivation rec {
  pname = "elpa";
  version = "2022.05.001";

  passthru = { inherit (blas) isILP64; };

  src = fetchurl {
    url = "https://elpa.mpcdf.mpg.de/software/tarball-archive/Releases/${version}/elpa-${version}.tar.gz";
    sha256 = "sha256-IH5vJtZTL7cDc6/D7z04JVITr2He9lnCXa06MOT8o4s=";
  };

  patches = [
    # Use a plain name for the pkg-config file
    ./pkg-config.patch
  ];

  postPatch = ''
    patchShebangs ./fdep/fortran_dependencies.pl
    patchShebangs ./test-driver

    # Fix the test script generator
    substituteInPlace Makefile.am --replace '#!/bin/bash' '#!${stdenv.shell}'
  '';

  nativeBuildInputs = [ autoreconfHook perl openssh ];

  buildInputs = [ mpi blas lapack scalapack ]
    ++ lib.optional enableCuda cudatoolkit;

  preConfigure = ''
    export FC="mpifort"
    export CC="mpicc"
    export CXX="mpicxx"
    export CPP="cpp"

    # These need to be set for configure to succeed
    export FCFLAGS="${lib.optionalString stdenv.hostPlatform.isx86_64 "-msse3 "
      + lib.optionalString avxSupport "-mavx "
      + lib.optionalString avx2Support "-mavx2 -mfma "
      + lib.optionalString avx512Support "-mavx512"}"

    export CFLAGS=$FCFLAGS
  '';

  configureFlags = [
    "--with-mpi"
    "--enable-openmp"
    "--without-threading-support-check-during-build"
  ] ++ lib.optional blas.isILP64 "--enable-64bit-integer-math-support"
    ++ lib.optional (!avxSupport) "--disable-avx"
    ++ lib.optional (!avx2Support) "--disable-avx2"
    ++ lib.optional (!avx512Support) "--disable-avx512"
    ++ lib.optional (!stdenv.hostPlatform.isx86_64) "--disable-sse"
    ++ lib.optional stdenv.hostPlatform.isx86_64 "--enable-sse-assembly"
    ++ lib.optionals enableCuda [  "--enable-nvidia-gpu" "--with-NVIDIA-GPU-compute-capability=${nvidiaArch}" ];

  doCheck = true;

  preCheck = ''
    #patchShebangs ./

    # make sure the test starts even if we have less than 4 cores
    export OMPI_MCA_rmaps_base_oversubscribe=1

    # Fix to make mpich run in a sandbox
    export HYDRA_IFACE=lo

    # Run dual threaded
    export OMP_NUM_THREADS=2

    # Reduce test problem sizes
    export TEST_FLAGS="1500 50 16"
  '';

  meta = with lib; {
    description = "Eigenvalue Solvers for Petaflop-Applications";
    homepage = "https://elpa.mpcdf.mpg.de/";
    license = licenses.lgpl3Only;
    platforms = platforms.linux;
    maintainers = [ maintainers.markuskowa ];
  };
}