xref: /petsc/config/examples/arch-alcf-polaris.py (revision 1166ac79e6ba74581d83280f431d2068845298ec)
15d94af91SJunchao Zhang#!/usr/bin/python3
25d94af91SJunchao Zhang
33ab125cbSJunchao Zhang# Use GNU compilers:
45d94af91SJunchao Zhang#
55d94af91SJunchao Zhang# Note cray-libsci provides BLAS etc. In summary, we have
6*1166ac79SJunchao Zhang# module use /soft/modulefiles
7*1166ac79SJunchao Zhang# module unload darshan
8*1166ac79SJunchao Zhang# module load cudatoolkit-standalone/12.4.1 PrgEnv-gnu cray-libsci
93ab125cbSJunchao Zhang#
105d94af91SJunchao Zhang# $ module list
115d94af91SJunchao Zhang# Currently Loaded Modules:
12*1166ac79SJunchao Zhang#   1) libfabric/1.15.2.0       6) nghttp2/1.57.0-ciat5hu         11) cray-dsmml/0.2.2    16) craype-x86-milan
13*1166ac79SJunchao Zhang#   2) craype-network-ofi       7) curl/8.4.0-2ztev25             12) cray-mpich/8.1.28   17) PrgEnv-gnu/8.5.0
14*1166ac79SJunchao Zhang#   3) perftools-base/23.12.0   8) cmake/3.27.7                   13) cray-pmi/6.1.13     18) cray-libsci/23.12.5
15*1166ac79SJunchao Zhang#   4) gcc-native/12.3          9) cudatoolkit-standalone/12.4.1  14) cray-pals/1.3.4
16*1166ac79SJunchao Zhang#   5) spack-pe-base/0.6.1     10) craype/2.7.30                  15) cray-libpals/1.3.4
175d94af91SJunchao Zhang
185d94af91SJunchao Zhangif __name__ == '__main__':
195d94af91SJunchao Zhang  import sys
205d94af91SJunchao Zhang  import os
215d94af91SJunchao Zhang  sys.path.insert(0, os.path.abspath('config'))
225d94af91SJunchao Zhang  import configure
235d94af91SJunchao Zhang  configure_options = [
245d94af91SJunchao Zhang    '--with-cc=cc',
255d94af91SJunchao Zhang    '--with-cxx=CC',
265d94af91SJunchao Zhang    '--with-fc=ftn',
275d94af91SJunchao Zhang    '--with-debugging=0',
285d94af91SJunchao Zhang    '--with-cuda',
295d94af91SJunchao Zhang    '--with-cudac=nvcc',
305d94af91SJunchao Zhang    '--with-cuda-arch=80', # Since there is no easy way to auto-detect the cuda arch on the gpu-less Polaris login nodes, we explicitly set it.
315d94af91SJunchao Zhang    '--download-kokkos',
325d94af91SJunchao Zhang    '--download-kokkos-kernels',
33*1166ac79SJunchao Zhang    '--download-hypre',
345d94af91SJunchao Zhang  ]
355d94af91SJunchao Zhang  configure.petsc_configure(configure_options)
365d94af91SJunchao Zhang
373ab125cbSJunchao Zhang# Use NVHPC compilers
383ab125cbSJunchao Zhang#
393ab125cbSJunchao Zhang# Unset so that cray won't add -gpu to nvc even when craype-accel-nvidia80 is loaded
403ab125cbSJunchao Zhang# unset CRAY_ACCEL_TARGET
413ab125cbSJunchao Zhang# module load nvhpc/22.11 PrgEnv-nvhpc
423ab125cbSJunchao Zhang#
433ab125cbSJunchao Zhang# I met two problems with nvhpc and Kokkos (and Kokkos-Kernels) 4.2.0.
443ab125cbSJunchao Zhang# 1) Kokkos-Kernles failed at configuration to find TPL cublas and cusparse from NVHPC.
453ab125cbSJunchao Zhang#    As a workaround, I just load cudatoolkit-standalone/11.8.0 to let KK use cublas and cusparse from cudatoolkit-standalone.
463ab125cbSJunchao Zhang# 2) KK failed at compilation
473ab125cbSJunchao Zhang# "/home/jczhang/petsc/arch-kokkos-dbg/externalpackages/git.kokkos-kernels/batched/dense/impl/KokkosBatched_Gemm_Serial_Internal.hpp", line 94: error: expression must have a constant value
483ab125cbSJunchao Zhang#     constexpr int nbAlgo = Algo::Gemm::Blocked::mb();
493ab125cbSJunchao Zhang#                            ^
503ab125cbSJunchao Zhang# "/home/jczhang/petsc/arch-kokkos-dbg/externalpackages/git.kokkos-kernels/blas/impl/KokkosBlas_util.hpp", line 58: note: cannot call non-constexpr function "__builtin_is_device_code" (declared implicitly)
513ab125cbSJunchao Zhang#           KOKKOS_IF_ON_HOST((return 4;))
523ab125cbSJunchao Zhang#           ^
533ab125cbSJunchao Zhang#           detected during:
543ab125cbSJunchao Zhang#
553ab125cbSJunchao Zhang# It is a KK problem and I have to wait for their fix.
56