Skip to content

Commit

Permalink
Upgrade ascent build system and use [email protected] on CI platforms (#20)
Browse files Browse the repository at this point in the history
* Boilerplate scripts to install modules on Ascent via submodule Spack

* Fix '--nnodes'-->'-nodes' on Ascent

* Improve Ascent env.sh

* [email protected] on Ascent

* Apply pre-commmit fixes

* Relax constraints on exago dependencies on Ascent and build ~python

* concretizer: reuse was causing several packages to be duplicated in the environment. Require clean concretizations on  Ascent.

* Minor module update on Ascent

* Add LAPACK_LIBRARIES to Ascent base script. CMAKE was picking up python's openblas otherwise.

* Error with unzip.

* Apply pre-commmit fixes

* Add working build on ascent.

* Add working gcc11.2.0 spack spec.

* Add Ascent Spack pipeline. [ascent-rebuild]

* Update gcc version to 11.2.0 in base.sh [skip-ci]

* Fix stages of Ascent pipeline [ascent-rebuild]

* Add working ascent spack build.

* Add hiop@develop force rebuild to PNNL CI [ascent-rebuild] [newell-rebuild] [deception-rebuild] [incline-rebuild].

* Update Ascent spack built tcl modules

* Only test ascent on tcl module update [ci-skip]

* Update base.sh to disable python on ascent [skip ci]

* Remove LAPACK_LIBRARIES spec [ascent-test]

* Update ascent.gitlab-ci.yml to fix needs/dependencies [ascent-test]

* Update deception spack built tcl modules - [deception-test]

* Try again with Python, but have Spack build it instead of using the external module [ascent-rebuild]

* Force python rebuild on ascent and use [email protected] on incline [ascent-rebuild] [newell-rebuild] [incline-rebuild]

* Pin [email protected] on all CI platforms [decetpion-rebuild] [ascent-rebuild] [newell-rebuild] [incline-rebuild]

* Fix false positive/negative in Ascent pipelines [deception-rebuild] [ascent-test]

* Update incline spack built tcl modules - [incline-test]

* Update newell spack built tcl modules - [newell-test]

* Fix HiOp spec on Ascent [ascent-rebuild].

* Update deception spack built tcl modules - [deception-test]

* Update CPU Spack build with issue for each failing build [ci skip]

* Update Ascent spack built tcl modules [ascent-test]

* Add 1.0.0 dep into CHANGELOG.

* Add ascent-skip to CI to get tests passing [ascent-test]

---------

Co-authored-by: nkoukpaizan <[email protected]>
Co-authored-by: Cameron Rutherford <[email protected]>
Co-authored-by: cameronrutherford <[email protected]>
Co-authored-by: spack-auto-module <[email protected]>
  • Loading branch information
5 people committed Oct 25, 2023
1 parent 136f43e commit 33143f3
Show file tree
Hide file tree
Showing 28 changed files with 656 additions and 413 deletions.
4 changes: 2 additions & 2 deletions .github/workflows/spack_cpu_build.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -12,8 +12,8 @@ jobs:
strategy:
matrix:
spack_spec:
# Minimal Build
# See #18
# Minimal Build(s)
# See #39 - ~python~mpi causes issues
# - exago@develop~mpi~ipopt~hiop~python~raja
- exago@develop+mpi~ipopt+hiop~python~raja ^openmpi
# See #18 - +hiop~mpi causes issues
Expand Down
2 changes: 2 additions & 0 deletions .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -12,3 +12,5 @@ node_modules
.env
package-lock.json
config.py
spack_*
hiop/
55 changes: 50 additions & 5 deletions .gitlab/ornl/ascent.gitlab-ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -102,6 +102,7 @@
- rm -rf "$WORKDIR"

Ascent Build:
needs: []
stage: build
tags:
- nobatch
Expand All @@ -112,27 +113,71 @@ Ascent Build:

Ascent Test:
stage: test
dependencies:
- 'Ascent Build'
needs: ['Ascent Build']
tags:
- batch
variables:
# Don't clone for test jobs
GIT_STRATEGY: none
CTESTARGS: "--output-on-failure"
CTESTARGS: "--output-on-failure -LE ascent-skip "
extends:
- .ascent_rules
- .ascent_test_script_template
- .ascent_environment_template

# -- Spack Module Rebuilding

Ascent-Module-Init:
needs: []
extends: .ascent_environment_template
variables:
GIT_SUBMODULE_STRATEGY: recursive
rules:
- if: '$CI_COMMIT_TITLE =~ /\[ascent-rebuild\]/'
when: always
stage: build
tags:
- nobatch
script:
- set -xv
- mkdir -p "$WORKDIR"
- cp -r . "$WORKDIR"
- cd "$WORKDIR"
- MY_CLUSTER=ascent ./buildsystem/spack/binary_mirror.sh
- res=$?
- exit $res

Ascent-Module-Build:
needs: ["Ascent-Module-Init"]
rules:
- if: '$CI_COMMIT_TITLE =~ /\[ascent-rebuild\]/'
when: always
extends: .ascent_environment_template
stage: test
tags:
- batch
script:
- cd "$WORKDIR"
- git checkout -b ${CI_COMMIT_REF_NAME}
- ./buildsystem/spack/ascent/bsub.sh
- git add ./buildsystem/spack/ascent/modules/*
- git config user.name "spack-auto-module"
- git config user.email "[email protected]"
- git commit -m "Update Ascent spack built tcl modules [ascent-test]"
- git remote set-url origin https://gitlab-ci-token:${SPACK_GIT_TOKEN}@github.com/pnnl/ExaGO.git
- git pull --rebase origin ${CI_COMMIT_REF_NAME}
- git push origin ${CI_COMMIT_REF_NAME}
after_script:
- cd "$WORKDIR/.."
- rm -rf "$WORKDIR"

# -- Reporting Ascent Status to PNNL GitHub
.report-job:
when: always
variables:
MY_CLUSTER: Ascent
GIT_STRATEGY: none
STATUS_PROJECT: pnnl/ExaGO
STATUS_NAME: ppc64le/gcc@10.2/cuda@11.4.2/v100@70
STATUS_NAME: ppc64le/gcc@11.2/cuda@11.8/v100@70
tags:
- nobatch
extends: .ascent_environment_template
Expand Down
3 changes: 3 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,9 +8,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [develop]

### General
- Transition from GitLab to GitHub
- ExaGO now depends on HiOp 1.0.0. Next release will require HiOp 1.0.1+

### Build system
- Add CI on Incline with Spack module pipelines (AMD / HIP GPU machine at PNNL)
- Add Spack module pipelines to Ascent

### PS

Expand Down
17 changes: 17 additions & 0 deletions buildsystem/gcc-cuda/ascent/base.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,17 @@
#!/bin/bash

export MY_CLUSTER=ascent

module reset

# Load system modules
module load gcc/11.2.0
module load spectrum-mpi/10.4.0.3-20210112

export CC=$(which gcc)
export CXX=$(which g++)
export FC=$(which gfortran)

export MY_PETSC_DIR=$PETSC_DIR
export EXTRA_CMAKE_ARGS="$EXTRA_CMAKE_ARGS -DEXAGO_CTEST_LAUNCH_COMMAND='jsrun -g 1 -n 1'"
export EXTRA_CMAKE_ARGS="$EXTRA_CMAKE_ARGS -DEXAGO_ENABLE_PYTHON=OFF"
75 changes: 0 additions & 75 deletions buildsystem/gcc-cuda/ascent/spack.yaml

This file was deleted.

128 changes: 6 additions & 122 deletions buildsystem/gcc-cuda/ascentVariables.sh
Original file line number Diff line number Diff line change
@@ -1,125 +1,9 @@
export MY_CLUSTER=ascent
#!/bin/bash

module purge
export SRCDIR=${SRCDIR:-$PWD}

module use -a /gpfs/wolf/proj-shared/csc359/src/spack/share/spack/modules/linux-rhel8-power9le
# Shared system configuration
source $SRCDIR/buildsystem/gcc-cuda/ascent/base.sh

# Load spack modules
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-blt/0.4.1/gcc-10.2.0-ndtzigt
# [email protected]%[email protected]+cuda~ipo~rocm~tests build_type=RelWithDebInfo cuda_arch=70 arch=linux-rhel8-power9le
module load exasgd-camp/0.2.2/cuda-11.4.2/gcc-10.2.0-2xxjut5
# [email protected]%[email protected]~doc+ncurses+ownlibs~qt build_type=Release arch=linux-rhel8-power9le
module load exasgd-cmake/3.22.2/gcc-10.2.0-sj55k3d
# [email protected]%[email protected]+blas arch=linux-rhel8-power9le
module load exasgd-coinhsl/2015.06.23/gcc-10.2.0-zr37tzv
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-cub/1.16.0/gcc-10.2.0-zeukz4z
# [email protected]%[email protected]~allow-unsupported-compilers~dev arch=linux-rhel8-power9le
module load exasgd-cuda/11.4.2/gcc-10.2.0-yf564wn
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-diffutils/3.8/gcc-10.2.0-n2gttub
# [email protected]_experimental%[email protected]+cuda~develtools~full_optimizations~hwloc~ipo~oneapi+openmp~rocm+shared build_system=cmake build_type=Debug cuda_arch=70 dev_path=/gpfs/wolf/proj-shared/csc359/src/ginkgo arch=linux-rhel8-power9le
module load exasgd-ginkgo/1.5.0.glu_experimental/cuda-11.4.2/gcc-10.2.0-h4v5snw
# [email protected]%[email protected] libs=shared,static arch=linux-rhel8-power9le
module load exasgd-gmp/6.2.1/gcc-10.2.0-6oxkdwz
# gnuconfig@2021-08-14%[email protected] arch=linux-rhel8-power9le
module load exasgd-gnuconfig/2021-08-14/gcc-10.2.0-j2pelq3
# [email protected]%[email protected]~cxx~fortran~hl~ipo~java+mpi+shared~szip~threadsafe+tools api=default build_type=RelWithDebInfo arch=linux-rhel8-power9le
module load exasgd-hdf5/1.12.2/spectrum-mpi-10.4.0.3-20210112/gcc-10.2.0-yh642ea
# hiop@develop%[email protected]+cuda+cusolver+deepchecking+ginkgo~ipo~jsrun+kron+mpi+raja~rocm~shared+sparse build_system=cmake build_type=RelWithDebInfo cuda_arch=70 dev_path=/gpfs/wolf/proj-shared/csc359/src/nicholson-hiop arch=linux-rhel8-power9le
module load exasgd-hiop/develop/cuda-11.4.2/spectrum-mpi-10.4.0.3-20210112/gcc-10.2.0-ucuclxs
# [email protected]%[email protected]~complex~cuda~debug+fortran~gptune~int64~internal-superlu~mixedint+mpi~openmp~rocm+shared~superlu-dist~unified-memory arch=linux-rhel8-power9le
module load exasgd-hypre/2.24.0/spectrum-mpi-10.4.0.3-20210112/gcc-10.2.0-oybqs3s
# [email protected]%[email protected]+coinhsl~debug~metis~mumps build_system=autotools arch=linux-rhel8-power9le
module load exasgd-ipopt/3.12.10/gcc-10.2.0-i4gqiqg
# [email protected]%[email protected] libs=shared,static arch=linux-rhel8-power9le
module load exasgd-libiconv/1.16/gcc-10.2.0-cpn6euq
# [email protected]%[email protected]+cuda+fortran~ipo~rocm+shared build_type=RelWithDebInfo cuda_arch=70 arch=linux-rhel8-power9le
module load exasgd-magma/2.6.2/cuda-11.4.2/gcc-10.2.0-ffujy7o
# [email protected]%[email protected]~gdb~int64~real64+shared build_type=Release patches=4991da9,b1225da arch=linux-rhel8-power9le
module load exasgd-metis/5.1.0/gcc-10.2.0-kdt6x4i
# [email protected]%[email protected] libs=shared,static arch=linux-rhel8-power9le
module load exasgd-mpfr/4.1.0/gcc-10.2.0-ua5zak3
# [email protected]%[email protected]~bignuma~consistent_fpcsr~ilp64+locking+pic+shared symbol_suffix=none threads=none arch=linux-rhel8-power9le
module load exasgd-openblas/0.3.20/gcc-10.2.0-lkjyfws
# [email protected]%[email protected]~gdb~int64~ipo+shared build_type=RelWithDebInfo patches=4f89253,50ed208,704b84f arch=linux-rhel8-power9le
module load exasgd-parmetis/4.0.3/spectrum-mpi-10.4.0.3-20210112/gcc-10.2.0-gqz32yk
# [email protected]%[email protected]~X~batch~cgns~complex~cuda~debug+double~exodusii~fftw+fortran~giflib+hdf5~hpddm~hwloc+hypre~int64~jpeg~knl~kokkos~libpng~libyaml~memkind+metis~mkl-pardiso~mmg~moab~mpfr+mpi~mumps~openmp~p4est~parmmg~ptscotch~random123~rocm~saws~scalapack+shared~strumpack~suite-sparse+superlu-dist~tetgen~trilinos~valgrind build_system=generic clanguage=C arch=linux-rhel8-power9le
module load exasgd-petsc/3.18.1/spectrum-mpi-10.4.0.3-20210112/gcc-10.2.0-cwok3k4
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-pkgconf/1.8.0/gcc-10.2.0-nngfigd
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-attrs/21.4.0/gcc-10.2.0-av2tzhn
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-flit-core/3.6.0/gcc-10.2.0-wdoqgro
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-importlib-metadata/4.8.2/gcc-10.2.0-klg3xix
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-iniconfig/1.1.1/gcc-10.2.0-o6bmfak
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-mpi4py/3.1.2/spectrum-mpi-10.4.0.3-20210112/gcc-10.2.0-q53tcvd
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-packaging/21.3/gcc-10.2.0-ej3e6zm
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-pip/21.3.1/gcc-10.2.0-awkoczv
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-pluggy/1.0.0/gcc-10.2.0-tlqdjwe
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-py/1.11.0/gcc-10.2.0-fwbtxbd
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-pyparsing/3.0.6/gcc-10.2.0-qxfb6l5
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-pytest/6.2.5/gcc-10.2.0-2vv5hdj
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-setuptools/59.4.0/gcc-10.2.0-elx3acw
# [email protected]%[email protected]+toml arch=linux-rhel8-power9le
module load exasgd-py-setuptools-scm/6.3.2/gcc-10.2.0-xdvaqbm
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-toml/0.10.2/gcc-10.2.0-6lhs4ix
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-tomli/1.2.2/gcc-10.2.0-4glf4eg
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-typing-extensions/4.1.1/gcc-10.2.0-qg6k6g2
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-wheel/0.37.0/gcc-10.2.0-tg3twdl
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-py-zipp/3.6.0/gcc-10.2.0-4vxciug
# [email protected]%[email protected]+bz2+ctypes+dbm~debug+ensurepip+libxml2+lzma~nis~optimizations+pic+pyexpat+pythoncmd+readline+shared+sqlite3+ssl~tix~tkinter~ucs4+uuid+zlib patches=c129e34 arch=linux-rhel8-power9le
module load exasgd-python/3.6.8/gcc-10.2.0-273robh
# [email protected]%[email protected]+cuda~examples~exercises~ipo+openmp~rocm+shared~tests build_type=RelWithDebInfo cuda_arch=70 arch=linux-rhel8-power9le
module load exasgd-raja/0.14.0/cuda-11.4.2/gcc-10.2.0-hkkig4v
# [email protected]%[email protected] arch=linux-rhel8-power9le
module load exasgd-spectrum-mpi/10.4.0.3-20210112/gcc-10.2.0-c5uqeqr
# [email protected]%[email protected]~cuda~graphblas~openmp+pic~tbb arch=linux-rhel8-power9le
module load exasgd-suite-sparse/5.10.1/gcc-10.2.0-sq5isrj
# [email protected]%[email protected]~cuda~int64~ipo~openmp~rocm+shared build_type=RelWithDebInfo patches=8da9e21 arch=linux-rhel8-power9le
module load exasgd-superlu-dist/7.2.0/spectrum-mpi-10.4.0.3-20210112/gcc-10.2.0-7cwmhkg
# [email protected]%[email protected]~c+cuda~device_alloc~deviceconst~examples~fortran~ipo~numa~openmp~rocm~shared build_type=RelWithDebInfo cuda_arch=70 tests=none arch=linux-rhel8-power9le
module load exasgd-umpire/6.0.0/cuda-11.4.2/gcc-10.2.0-selkipw
# [email protected]%[email protected]+optimize+pic+shared patches=0d38234 arch=linux-rhel8-power9le
module load exasgd-zlib/1.2.12/gcc-10.2.0-lfmc24z

# Load system modules
module load cuda/11.4.2
module load gcc/10.2.0
module load spectrum-mpi/10.4.0.3-20210112
module load cmake/3.22.2

export MY_PETSC_DIR=$PETSC_DIR
export EXTRA_CMAKE_ARGS="$EXTRA_CMAKE_ARGS -DEXAGO_CTEST_LAUNCH_COMMAND='jsrun -g 1 -n 1'"

[ -f $PWD/nvblas.conf ] && rm $PWD/nvblas.conf
cat > $PWD/nvblas.conf <<-EOD
NVBLAS_LOGFILE nvblas.log
NVBLAS_CPU_BLAS_LIB $OPENBLAS_LIBRARY_DIR/libopenblas.so
NVBLAS_GPU_LIST ALL
NVBLAS_TILE_DIM 2048
NVBLAS_AUTOPIN_MEM_ENABLED
EOD
export NVBLAS_CONFIG_FILE=$PWD/nvblas.conf
echo "Generated $PWD/nvblas.conf"

export CC=$(which gcc)
export CXX=$(which g++)
export FC=$(which gfortran)
# Spack modules
source $SRCDIR/buildsystem/spack/$MY_CLUSTER/modules/dependencies.sh
Loading

0 comments on commit 33143f3

Please sign in to comment.