diff --git a/.ipr/hibench.job b/.ipr/hibench.job index 70c96b2..0838464 100644 --- a/.ipr/hibench.job +++ b/.ipr/hibench.job @@ -8,16 +8,22 @@ code_version="$4" # git branch id or commit id eg : 'a3bed1c3ccfbca572003020d3e cmake_options="$5" # eg '-DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=ON' benchmark_command="$6" # eg 'ctest -L ^arch4_quick$' -if [ "${JOBID}" = '' ] +cmake_path='/opt/cmake/cmake-3.23.0/bin/cmake' + +if [ "${JOB_ID}" = '' ] then # this script is not executed by sge... set dummy values for test - TMP_DIR=/tmp - JOBID=666666 + TMPDIR=/tmp + JOB_ID=666666 NSLOTS=2 fi -echo "Executing on $(hostname)" -temp_dir=${TMP_DIR}/$(whoami)/${JOBID} +echo "Executing job ${JOB_ID} on $(hostname)" +temp_dir=${TMPDIR}/$(whoami)/${JOB_ID} +if [ -d "${temp_dir}" ] +then + rm -Rf "${temp_dir}" +fi mkdir -p "${temp_dir}" starbench_path="${temp_dir}/starbench.py" @@ -38,6 +44,7 @@ command="${command} --git-pass-file ${git_pass_file}" command="${command} --num-cores ${num_cores}" command="${command} --output-dir ${output_dir}" command="${command} --code-version ${code_version}" +command="${command} --cmake-path ${cmake_path}" # echo "cmake_options: @$cmake_options@" for cmake_option in ${cmake_options} do diff --git a/.ipr/launch-perf-jobs.sh b/.ipr/launch-perf-jobs.sh index 8a956b6..ff854d3 100755 --- a/.ipr/launch-perf-jobs.sh +++ b/.ipr/launch-perf-jobs.sh @@ -1,6 +1,23 @@ #!/usr/bin/env bash # this script launches jobs to run hibridon benchmarks on physix cluster for the given version of hibridon (commit number) +function show_usage() +{ + echo "launches hibridon benchmark jobs on IPR's physix cluster" + echo + echo "syntax :" + echo " $0 " + echo + echo "example:" + echo " $0 a3bed1c3ccfbca572003020d3e3d3b1ff3934fad" +} + +if [ "$#" != "1" ] +then + show_usage + exit 1 +fi + HIBRIDON_VERSION="$1" # the version of hibridon to test, in the form of a valid commit number eg 'a3bed1c3ccfbca572003020d3e3d3b1ff3934fad' # 'a3bed1c3ccfbca572003020d3e3d3b1ff3934fad' # latest from branch master as of 01/06/2022 12:52 # code_version='775048db02dfb317d5eaddb6d6db520be71a2fdf' # latest from branch graffy-issue51 as of 01/06/ @@ -21,36 +38,59 @@ function launch_job_for_host_group() local hosts='' local num_cores='' case "${host_group_id}" in - 'xeon_gold_6140') - hosts='\ - physix88.ipr.univ-rennes1.fr|\ - physix89.ipr.univ-rennes1.fr' + 'intel_xeon_gold_6140') + hosts="${hosts}physix88.ipr.univ-rennes1.fr" + hosts="${hosts}|physix89.ipr.univ-rennes1.fr" num_cores='36' ;; + 'intel_xeon_x5550') + hosts="${hosts}physix48.ipr.univ-rennes1.fr" + num_cores='8' + ;; *) error "unhandled host_group_id : ${host_group_id}" exit 1 ;; esac - quick_test='^arch4_quick$' # about 2s on a core i5 8th generation - benchmark_test='^nh3h2_qma_long$' # about 10min on a core i5 8th generation + quick_test='arch4_quick' # about 2s on a core i5 8th generation + representative_test='nh3h2_qma_long' # about 10min on a core i5 8th generation + benchmark_test="${quick_test}" git_repos_url="https://github.com/hibridon/hibridon" git_user='g-raffy' # os.environ['HIBRIDON_REPOS_USER'] git_pass_file="$HOME/.github/personal_access_tokens/bench.hibridon.cluster.ipr.univ-rennes1.fr.pat" cmake_options='-DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=ON' - benchmark_command="ctest --output-on-failure -L ${quick_test}" + benchmark_command="ctest --output-on-failure -L ^${benchmark_test}\$" # cat $SCRIPT_DIR/hibench.job | sed "s~~$(cat $SCRIPT_DIR/starbench.py)~" > /tmp/hibench.job cat $SCRIPT_DIR/hibench.job | substitute_TAG_with_FILEcontents '' "$SCRIPT_DIR/starbench.py" > /tmp/hibench.job chmod a+x /tmp/hibench.job - command="/tmp/hibench.job \"${git_repos_url}\" \"${git_user}\" \"${git_pass_file}\" \"${HIBRIDON_VERSION}\" \"${cmake_options}\" \"${benchmark_command}\"" - echo "command = $command" - eval $command + local hibench_root_dir="$GLOBAL_WORK_DIR/graffy/hibridon/benchmarks/starbench" + mkdir -p "${hibench_root_dir}" - # qsub -pe smp "$num_cores" -l "hostname=${hosts}" + local this_bench_dir="${hibench_root_dir}/${hibridon_version}/${benchmark_test}/${host_group_id}" + mkdir -p "${this_bench_dir}" + + command="/tmp/hibench.job \"${git_repos_url}\" \"${git_user}\" \"${git_pass_file}\" \"${hibridon_version}\" \"${cmake_options}\" \"${benchmark_command}\"" + echo "command = $command" + # eval $command + + pushd "${this_bench_dir}" + + qsub_command="qsub" + qsub_command="${qsub_command} -pe smp ${num_cores}" + qsub_command="${qsub_command} -l \"hostname=${hosts}\"" + qsub_command="${qsub_command} -cwd" + qsub_command="${qsub_command} -m ae" + qsub_command="${qsub_command} -l mem_available=1G" + qsub_command="${qsub_command} -N ${benchmark_test}_${host_group_id}" + qsub_command="${qsub_command} ${command}" + # qsub -pe smp "$num_cores" -l "hostname=${hosts}" + echo "qsub_command = $qsub_command" + eval $qsub_command + popd } @@ -58,7 +98,8 @@ function launch_perf_jobs() { local hibridon_version="$1" # the version of hibridon to test, in the form of a valid commit number eg 'a3bed1c3ccfbca572003020d3e3d3b1ff3934fad' - launch_job_for_host_group "${hibridon_version}" 'xeon_gold_6140' + launch_job_for_host_group "${hibridon_version}" 'intel_xeon_x5550' + #launch_job_for_host_group "${hibridon_version}" 'intel_xeon_gold_6140' } diff --git a/.ipr/starbench.py b/.ipr/starbench.py index 82d557c..92933af 100755 --- a/.ipr/starbench.py +++ b/.ipr/starbench.py @@ -186,13 +186,15 @@ class StarBencher(): self._runs[run.id] = run def run(self): + print("executing the following command in parallel (%d parallel runs) : '%s'" % (self.num_parallel_runs, str(self.run_command))) for worker_id in range(self.num_parallel_runs): self._start_run(worker_id) # wait until all runs have finished self._finished_event.wait() with self._runs_lock: - if not all([run.return_code == 0 for run in self._runs.values()]): - raise Exception('at least one run failed') + workers_success = [run.return_code == 0 for run in self._runs.values()] + if not all(workers_success): + raise Exception('at least one run failed (workers_success = %s)' % workers_success) mean_duration, num_runs = self._get_run_mean_duration() print('mean duration : %.3f s (%d runs)' % (mean_duration, num_runs)) return mean_duration @@ -214,7 +216,7 @@ def test_starbencher(): # end of starbencher -def starbench_cmake_app(git_repos_url: str, code_version: str, tmp_dir: Path, num_cores: int, git_user: str, git_password: str, benchmark_command: List[str], cmake_options: List[str] = None): +def starbench_cmake_app(git_repos_url: str, code_version: str, tmp_dir: Path, num_cores: int, git_user: str, git_password: str, benchmark_command: List[str], cmake_options: List[str] = None, cmake_exe_location: Path = None): """ tests_to_run : regular expression as understood by ctest's -L option. eg '^arch4_quick$' """ @@ -247,8 +249,11 @@ def starbench_cmake_app(git_repos_url: str, code_version: str, tmp_dir: Path, nu # build_dir.mkdir(exist_ok=True) print('configuring %s into %s ...' % (src_dir, build_dir)) + cmake_prog = 'cmake' + if cmake_exe_location: + cmake_prog = str(cmake_exe_location) configure = StarBencher( - run_command=['cmake'] + cmake_options + [src_dir], + run_command=[cmake_prog] + cmake_options + [src_dir], num_cores_per_run=1, num_parallel_runs=num_cores, max_num_cores=num_cores, @@ -289,7 +294,7 @@ if __name__ == '__main__': example_text = '''example: - %(prog)s --git-repos-url https://github.com/hibridon/hibridon --code-version a3bed1c3ccfbca572003020d3e3d3b1ff3934fad --git-user g-raffy --git-pass-file "$HOME/.github/personal_access_tokens/bench.hibridon.cluster.ipr.univ-rennes1.fr.pat" --num-cores 2 --output-dir=/tmp/hibench --cmake-option=-DCMAKE_BUILD_TYPE=Release --cmake-option=-DBUILD_TESTING=ON --benchmark-command='ctest --output-on-failure -L ^arch4_quick$' + %(prog)s --git-repos-url https://github.com/hibridon/hibridon --code-version a3bed1c3ccfbca572003020d3e3d3b1ff3934fad --git-user g-raffy --git-pass-file "$HOME/.github/personal_access_tokens/bench.hibridon.cluster.ipr.univ-rennes1.fr.pat" --num-cores 2 --output-dir=/tmp/hibench --cmake-path=/opt/cmake/cmake-3.23.0/bin/cmake --cmake-option=-DCMAKE_BUILD_TYPE=Release --cmake-option=-DBUILD_TESTING=ON --benchmark-command='ctest --output-on-failure -L ^arch4_quick$' ''' @@ -302,6 +307,7 @@ if __name__ == '__main__': password_group.add_argument('--git-pass', type=str, help='the password (or personal access token) to use (not recommended for security reasons)') parser.add_argument('--num-cores', type=int, required=True, help='the number of cores that the benchmark will use') parser.add_argument('--output-dir', type=Path, required=True, help='where the output files will be placed') + parser.add_argument('--cmake-path', type=Path, help='the path to the cmake executable to use in case a specific cmake is wanted') parser.add_argument('--cmake-option', type=str, action='append', help='additional option passed to cmake in the configure step (use this flag multiple times if you need more than one cmake option)') parser.add_argument('--benchmark-command', required=True, type=str, help='the command to benchmark') args = parser.parse_args() @@ -316,4 +322,4 @@ if __name__ == '__main__': with open(args.git_pass_file, 'r') as f: git_password = f.readline().replace('\n', '') # os.environ['HIBRIDON_REPOS_PAT'] - starbench_cmake_app(git_repos_url=git_repos_url, code_version=args.code_version, tmp_dir=args.output_dir, num_cores=args.num_cores, git_user=git_user, git_password=git_password, cmake_options=args.cmake_option, benchmark_command=args.benchmark_command.split(' ')) + starbench_cmake_app(git_repos_url=git_repos_url, code_version=args.code_version, tmp_dir=args.output_dir, num_cores=args.num_cores, git_user=git_user, git_password=git_password, cmake_options=args.cmake_option, benchmark_command=args.benchmark_command.split(' '), cmake_exe_location=args.cmake_path)