From 4b6540b889cfce2c47801b3a4457691661d90557 Mon Sep 17 00:00:00 2001
From: "Christopher J. Morrone" <morrone2@llnl.gov>
Date: Mon, 15 Aug 2005 18:57:55 +0000
Subject: [PATCH] Add files that were missing from the r5239 commit
---
src/common/dist_tasks.c | 151 ++++++++++++++++++++++++++++++++++++++++
src/common/dist_tasks.h | 67 ++++++++++++++++++
2 files changed, 218 insertions(+)
create mode 100644 src/common/dist_tasks.c
create mode 100644 src/common/dist_tasks.h
diff --git a/src/common/dist_tasks.c b/src/common/dist_tasks.c
new file mode 100644
index 00000000000..a35f89a6d05
--- /dev/null
+++ b/src/common/dist_tasks.c
@@ -0,0 +1,151 @@
+/*****************************************************************************\
+ * dist_tasks.c - function to distribute tasks over nodes.
+ *****************************************************************************
+ *
+ * Copyright (C) 2005 Hewlett-Packard Development Company, L.P.
+ * Written by Chris Holmes, <cholmes@hp.com>, who borrowed heavily
+ * from other parts of SLURM.
+ *
+ * This file is part of SLURM, a resource management program.
+ * For details, see <http://www.llnl.gov/linux/slurm/>.
+ *
+ * SLURM is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
+ * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with SLURM; if not, write to the Free Software Foundation, Inc.,
+ * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+ *
+ * This file is patterned after hostlist.c, written by Mark Grondona and
+ * Copyright (C) 2002 The Regents of the University of California.
+\*****************************************************************************/
+
+#ifdef HAVE_CONFIG_H
+# include "config.h"
+# if HAVE_STRING_H
+# include <string.h>
+# endif
+#else /* !HAVE_CONFIG_H */
+# include <string.h>
+#endif /* HAVE_CONFIG_H */
+
+
+#include <stdlib.h>
+
+#include <slurm/slurm_errno.h>
+
+#include "src/common/dist_tasks.h"
+#include "src/common/hostlist.h"
+#include "src/common/log.h"
+#include "src/common/xmalloc.h"
+
+
+/*
+ * distribute_tasks - determine how many tasks of a job will be run on each.
+ * node. Distribution is influenced by number of cpus on
+ * each host.
+ * IN mlist - hostlist corresponding to cpu arrays
+ * IN num_cpu_groups - elements in below cpu arrays
+ * IN cpus_per_node - cpus per node
+ * IN cpu_count_reps - how many nodes have same cpu count
+ * IN tlist - hostlist of nodes on which to distribute tasks
+ * IN num_tasks - number of tasks to distribute across these cpus
+ * RET a pointer to an integer array listing task counts per node
+ * NOTE: allocates memory that should be xfreed by caller
+ */
+int *distribute_tasks(const char *mlist, uint16_t num_cpu_groups,
+ uint32_t *cpus_per_node, uint32_t *cpu_count_reps,
+ const char *tlist, uint32_t num_tasks)
+{
+ hostlist_t master_hl = NULL, task_hl = NULL;
+ int i, index, count, hostid, nnodes, ncpus, *cpus, *ntask = NULL;
+ char *this_node_name;
+
+ if (!tlist || num_tasks == 0)
+ return NULL;
+
+ if ((master_hl = hostlist_create(mlist)) == NULL)
+ fatal("hostlist_create error for %s: %m", mlist);
+
+ if ((task_hl = hostlist_create(tlist)) == NULL)
+ fatal("hostlist_create error for %s: %m", tlist);
+
+ nnodes = hostlist_count(task_hl);
+ ntask = (int *) xmalloc(sizeof(int *) * nnodes);
+ if (!ntask) {
+ slurm_seterrno(ENOMEM);
+ return NULL;
+ }
+
+ index = 0;
+ count = 1;
+ i = 0;
+ ncpus = 0;
+ while ((this_node_name = hostlist_shift(master_hl))) {
+
+ if (hostlist_find(task_hl, this_node_name) >= 0) {
+ if (i >= nnodes)
+ fatal("Internal error: duplicate nodes? (%s)(%s):%m", mlist, tlist);
+ ntask[i++] = cpus_per_node[index];
+ ncpus += cpus_per_node[index];
+ }
+
+ if (++count > cpu_count_reps[index]) {
+ index++;
+ count = 1;
+ }
+ free(this_node_name);
+ }
+ hostlist_destroy(task_hl);
+
+ if (num_tasks >= ncpus) {
+ /*
+ * Evenly overcommit tasks over the hosts
+ */
+ int extra = num_tasks - ncpus;
+ int add_to_all = extra / nnodes;
+ int subset = extra % nnodes;
+ for (i = 0; i < nnodes; i++) {
+ ntask[i] += add_to_all;
+ if (i < subset)
+ ntask[i]++;
+ }
+ return ntask;
+ }
+
+ /*
+ * NOTE: num_tasks is less than ncpus here.
+ *
+ * In a cyclic fashion, place tasks on the nodes as permitted
+ * by the cpu constraints.
+ */
+ cpus = ntask;
+ ntask = (int *) xmalloc(sizeof(int *) * nnodes);
+ if (!ntask) {
+ slurm_seterrno(ENOMEM);
+ xfree(cpus);
+ return NULL;
+ }
+
+ for (i = 0; i < nnodes; i++)
+ ntask[i] = 0;
+
+ hostid = 0;
+ for (i = 0; i < num_tasks;) {
+ if (ntask[hostid] < cpus[hostid]) {
+ ntask[hostid]++;
+ i++;
+ }
+ if (++hostid >= nnodes)
+ hostid = 0;
+ }
+ xfree(cpus);
+ return ntask;
+}
diff --git a/src/common/dist_tasks.h b/src/common/dist_tasks.h
new file mode 100644
index 00000000000..b347aebbea7
--- /dev/null
+++ b/src/common/dist_tasks.h
@@ -0,0 +1,67 @@
+/*****************************************************************************\
+ * dist_tasks.c - function to distribute tasks over nodes.
+ * $Id: slurm.hp.elan.patch,v 1.1 2005/07/28 04:08:19 cholmes Exp $
+ *****************************************************************************
+ *
+ * Copyright (C) 2005 Hewlett-Packard Development Company, L.P.
+ * Written by Chris Holmes, <cholmes@hp.com>, who borrowed heavily
+ * from other parts of SLURM.
+ *
+ * This file is part of SLURM, a resource management program.
+ * For details, see <http://www.llnl.gov/linux/slurm/>.
+ *
+ * SLURM is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
+ * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with SLURM; if not, write to the Free Software Foundation, Inc.,
+ * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+ *
+ * This file is patterned after hostlist.h, written by Mark Grondona and
+ * Copyright (C) 2002 The Regents of the University of California.
+\*****************************************************************************/
+
+#ifndef _DIST_TASKS_H
+#define _DIST_TASKS_H
+
+#if HAVE_CONFIG_H
+# include "config.h"
+# if HAVE_INTTYPES_H
+# include <inttypes.h>
+# else
+# if HAVE_STDINT_H
+# include <stdint.h>
+# endif
+# endif /* HAVE_INTTYPES_H */
+#endif
+
+
+/*
+ * distribute_tasks - determine how many tasks of a job will be run on each.
+ * node. Distribution is influenced by number of cpus on
+ * each host.
+ * IN mlist - hostlist corresponding to cpu arrays
+ * IN num_cpu_groups - elements in below cpu arrays
+ * IN cpus_per_node - cpus per node
+ * IN cpu_count_reps - how many nodes have same cpu count
+ * IN tlist - hostlist of nodes on which to distribute tasks
+ * (assumed to be a subset of masterlist)
+ * IN num_tasks - number of tasks to distribute across these cpus
+ * RET a pointer to an integer array listing task counts per node
+ * NOTE: allocates memory that should be xfreed by caller
+ */
+int * distribute_tasks(const char *mlist,
+ uint16_t num_cpu_groups,
+ uint32_t *cpus_per_node,
+ uint32_t *cpu_count_reps,
+ const char *tlist,
+ uint32_t num_tasks);
+
+#endif /* !_DIST_TASKS_H */
--
GitLab