From 70586691dae18cb94ee3e4c5366bd3d52f16b8d9 Mon Sep 17 00:00:00 2001
From: Danny Auble <da@llnl.gov>
Date: Tue, 22 Nov 2005 22:01:59 +0000
Subject: [PATCH] cyclic code taken out of federation plugin. functionality is
still there
---
src/plugins/switch/federation/federation.c | 159 +++++++-----------
src/plugins/switch/federation/federation.h | 4 +-
.../switch/federation/switch_federation.c | 2 +-
3 files changed, 60 insertions(+), 105 deletions(-)
diff --git a/src/plugins/switch/federation/federation.c b/src/plugins/switch/federation/federation.c
index 48b19b4f6aa..6ca50961efb 100644
--- a/src/plugins/switch/federation/federation.c
+++ b/src/plugins/switch/federation/federation.c
@@ -123,7 +123,6 @@ struct fed_jobinfo {
char job_desc[DESCLEN];
uint32_t window_memory;
uint8_t bulk_xfer; /* flag */
- uint8_t cyclic; /* flag */
uint16_t tables_per_task;
fed_tableinfo_t *tableinfo;
};
@@ -1579,7 +1578,11 @@ _job_step_window_state(fed_jobinfo_t *jp, hostlist_t hl, enum NTBL_RC state)
int nnodes;
int i, j;
int rc;
-
+ int task_cnt;
+ int full_node_cnt;
+ int min_procs_per_node;
+ int max_procs_per_node;
+
xassert(!hostlist_is_empty(hl));
xassert(jp);
xassert(jp->magic == FED_JOBINFO_MAGIC);
@@ -1598,57 +1601,35 @@ _job_step_window_state(fed_jobinfo_t *jp, hostlist_t hl, enum NTBL_RC state)
nprocs = jp->tableinfo[0].table_length;
hi = hostlist_iterator_create(hl);
- if (jp->cyclic) {
- _lock();
- for (proc_cnt = 0; proc_cnt < nprocs; proc_cnt++) {
- host = hostlist_next(hi);
- if(!host) {
- hostlist_iterator_reset(hi);
- host = hostlist_next(hi);
- }
+ debug("Allocating windows");
+ nnodes = hostlist_count(hl);
+ full_node_cnt = nprocs % nnodes;
+ min_procs_per_node = nprocs / nnodes;
+ max_procs_per_node = (nprocs + nnodes - 1) / nnodes;
+
+ proc_cnt = 0;
+ _lock();
+ for (i = 0; i < nnodes; i++) {
+ host = hostlist_next(hi);
+ if(!host)
+ error("Failed to get next host");
+
+ if(i < full_node_cnt)
+ task_cnt = max_procs_per_node;
+ else
+ task_cnt = min_procs_per_node;
+
+ for (j = 0; j < task_cnt; j++) {
rc = _window_state_set(jp->tables_per_task,
jp->tableinfo,
host, proc_cnt,
state);
- free(host);
+ proc_cnt++;
}
- _unlock();
- } else {
- int task_cnt;
- int full_node_cnt;
- int min_procs_per_node;
- int max_procs_per_node;
-
- debug("Allocating windows in block mode");
- nnodes = hostlist_count(hl);
- full_node_cnt = nprocs % nnodes;
- min_procs_per_node = nprocs / nnodes;
- max_procs_per_node = (nprocs + nnodes - 1) / nnodes;
-
- proc_cnt = 0;
- _lock();
- for (i = 0; i < nnodes; i++) {
- host = hostlist_next(hi);
- if(!host)
- error("Failed to get next host");
-
- if(i < full_node_cnt)
- task_cnt = max_procs_per_node;
- else
- task_cnt = min_procs_per_node;
-
- for (j = 0; j < task_cnt; j++) {
- rc = _window_state_set(jp->tables_per_task,
- jp->tableinfo,
- host, proc_cnt,
- state);
- proc_cnt++;
- }
- free(host);
- }
- _unlock();
+ free(host);
}
-
+ _unlock();
+
hostlist_iterator_destroy(hi);
return SLURM_SUCCESS;
}
@@ -1682,14 +1663,14 @@ fed_job_step_allocated(fed_jobinfo_t *jp, hostlist_t hl)
/* Setup everything for the job. Assign tasks across
- * nodes in a block or cyclic fashion and create the network table used
+ * nodes based on the hostlist given and create the network table used
* on all nodes of the job.
*
* Used by: slurmctld
*/
int
fed_build_jobinfo(fed_jobinfo_t *jp, hostlist_t hl, int nprocs,
- int cyclic, bool sn_all, int bulk_xfer)
+ bool sn_all, int bulk_xfer)
{
int nnodes;
hostlist_iterator_t hi;
@@ -1699,7 +1680,11 @@ fed_build_jobinfo(fed_jobinfo_t *jp, hostlist_t hl, int nprocs,
int i, j;
fed_nodeinfo_t *node;
int rc;
-
+ int task_cnt;
+ int full_node_cnt;
+ int min_procs_per_node;
+ int max_procs_per_node;
+
assert(jp);
assert(jp->magic == FED_JOBINFO_MAGIC);
assert(!hostlist_is_empty(hl));
@@ -1707,7 +1692,6 @@ fed_build_jobinfo(fed_jobinfo_t *jp, hostlist_t hl, int nprocs,
if(nprocs <= 0)
slurm_seterrno_ret(EINVAL);
- jp->cyclic = (uint8_t) cyclic;
jp->bulk_xfer = (uint8_t) bulk_xfer;
jp->job_key = _next_key();
snprintf(jp->job_desc, DESCLEN,
@@ -1746,17 +1730,24 @@ fed_build_jobinfo(fed_jobinfo_t *jp, hostlist_t hl, int nprocs,
}
}
-
- if (cyclic) {
- /* Allocate tables for one process per node at a time */
- debug("Allocating windows in cyclic mode");
- _lock();
- for (proc_cnt = 0; proc_cnt < nprocs; proc_cnt++) {
- host = hostlist_next(hi);
- if(!host) {
- hostlist_iterator_reset(hi);
- host = hostlist_next(hi);
- }
+ debug("Allocating windows");
+ nnodes = hostlist_count(hl);
+ full_node_cnt = nprocs % nnodes;
+ min_procs_per_node = nprocs / nnodes;
+ max_procs_per_node = (nprocs + nnodes - 1) / nnodes;
+ proc_cnt = 0;
+ _lock();
+ for (i = 0; i < nnodes; i++) {
+ host = hostlist_next(hi);
+ if(!host)
+ error("Failed to get next host");
+
+ if(i < full_node_cnt)
+ task_cnt = max_procs_per_node;
+ else
+ task_cnt = min_procs_per_node;
+
+ for (j = 0; j < task_cnt; j++) {
rc = _allocate_windows(jp->tables_per_task,
jp->tableinfo,
host, proc_cnt);
@@ -1764,46 +1755,12 @@ fed_build_jobinfo(fed_jobinfo_t *jp, hostlist_t hl, int nprocs,
_unlock();
goto fail;
}
- free(host);
+ proc_cnt++;
}
- _unlock();
- } else {
- int task_cnt;
- int full_node_cnt;
- int min_procs_per_node;
- int max_procs_per_node;
-
- debug("Allocating windows in non-cyclic mode");
- nnodes = hostlist_count(hl);
- full_node_cnt = nprocs % nnodes;
- min_procs_per_node = nprocs / nnodes;
- max_procs_per_node = (nprocs + nnodes - 1) / nnodes;
- proc_cnt = 0;
- _lock();
- for (i = 0; i < nnodes; i++) {
- host = hostlist_next(hi);
- if(!host)
- error("Failed to get next host");
-
- if(i < full_node_cnt)
- task_cnt = max_procs_per_node;
- else
- task_cnt = min_procs_per_node;
-
- for (j = 0; j < task_cnt; j++) {
- rc = _allocate_windows(jp->tables_per_task,
- jp->tableinfo,
- host, proc_cnt);
- if (rc != SLURM_SUCCESS) {
- _unlock();
- goto fail;
- }
- proc_cnt++;
- }
- free(host);
- }
- _unlock();
+ free(host);
}
+ _unlock();
+
#if FED_DEBUG
_print_table(jp->tableinfo[i].table, jp->tableinfo[i].table_length);
@@ -1848,7 +1805,6 @@ fed_pack_jobinfo(fed_jobinfo_t *j, Buf buf)
packmem(j->job_desc, DESCLEN, buf);
pack32(j->window_memory, buf);
pack8(j->bulk_xfer, buf);
- pack8(j->cyclic, buf);
pack16(j->tables_per_task, buf);
for (i = 0; i < j->tables_per_task; i++) {
_pack_tableinfo(&j->tableinfo[i], buf);
@@ -1903,7 +1859,6 @@ fed_unpack_jobinfo(fed_jobinfo_t *j, Buf buf)
goto unpack_error;
safe_unpack32(&j->window_memory, buf);
safe_unpack8(&j->bulk_xfer, buf);
- safe_unpack8(&j->cyclic, buf);
safe_unpack16(&j->tables_per_task, buf);
j->tableinfo = (fed_tableinfo_t *) xmalloc(j->tables_per_task
diff --git a/src/plugins/switch/federation/federation.h b/src/plugins/switch/federation/federation.h
index 9370cabef30..235e88eb167 100644
--- a/src/plugins/switch/federation/federation.h
+++ b/src/plugins/switch/federation/federation.h
@@ -73,8 +73,8 @@ int fed_pack_nodeinfo(fed_nodeinfo_t *np, Buf buf);
int fed_unpack_nodeinfo(fed_nodeinfo_t *np, Buf buf);
void fed_free_nodeinfo(fed_nodeinfo_t *np, bool ptr_into_array);
int fed_alloc_jobinfo(fed_jobinfo_t **jh);
-int fed_build_jobinfo(fed_jobinfo_t *jp, hostlist_t hl, int nprocs,
- int cyclic, bool sn_all, int bulk_xfer);
+int fed_build_jobinfo(fed_jobinfo_t *jp, hostlist_t hl, int nprocs,
+ bool sn_all, int bulk_xfer);
int fed_pack_jobinfo(fed_jobinfo_t *jp, Buf buf);
int fed_unpack_jobinfo(fed_jobinfo_t *jp, Buf buf);
fed_jobinfo_t *fed_copy_jobinfo(fed_jobinfo_t *jp);
diff --git a/src/plugins/switch/federation/switch_federation.c b/src/plugins/switch/federation/switch_federation.c
index 3c5798f4926..ad43a58d645 100644
--- a/src/plugins/switch/federation/switch_federation.c
+++ b/src/plugins/switch/federation/switch_federation.c
@@ -381,7 +381,7 @@ int switch_p_build_jobinfo(switch_jobinfo_t switch_job, char *nodelist,
|| strstr(network, "BULK_XFER"))
bulk_xfer = 1;
err = fed_build_jobinfo((fed_jobinfo_t *)switch_job, list,
- nprocs, cyclic_alloc, sn_all,
+ nprocs, sn_all,
bulk_xfer);
hostlist_destroy(list);
--
GitLab