/* Copyright (c) 2021-2023 Barcelona Supercomputing Center (BSC) * SPDX-License-Identifier: GPL-3.0-or-later */ #include #include #include #include #include #include #include static int ncpus = -1; static long nruns = 3L; static long ntasks = 200L; static atomic_int nhandles = 0; static void **handle; #pragma oss task static void do_task(int t) { handle[t] = nanos6_get_current_blocking_context(); if (handle[t] == NULL) abort(); atomic_fetch_add(&nhandles, 1); nanos6_block_current_task(handle[t]); } static void do_run(void) { memset(handle, 0, ntasks * sizeof(void *)); atomic_store(&nhandles, 0); for (int t = 0; t < ntasks; t++) do_task(t); /* Wait for all tasks to fill the handle */ while (atomic_load(&nhandles) < ntasks) ; /* Is ok if we call unblock before the block happens */ for (int t = 0; t < ntasks; t++) { if (handle[t] == NULL) abort(); nanos6_unblock_task(handle[t]); } #pragma oss taskwait } static int get_ncpus(void) { return (int) nanos6_get_num_cpus(); } int main(void) { ncpus = get_ncpus(); handle = calloc(ntasks, sizeof(void *)); if (handle == NULL) { perror("calloc failed"); return -1; } for (int run = 0; run < nruns; run++) do_run(); return 0; }