From a652bdf6f90ebef904d58c55db3b5a3fc0819ee5 Mon Sep 17 00:00:00 2001
From: Moe Jette <jette1@llnl.gov>
Date: Fri, 3 Apr 2009 02:11:39 +0000
Subject: [PATCH] Add logic to override task bindings and set more reasonable
limits in the event that the hardware does not match the configuration
information.
---
src/plugins/task/affinity/dist_tasks.c | 18 ++++++++++++++++--
1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/src/plugins/task/affinity/dist_tasks.c b/src/plugins/task/affinity/dist_tasks.c
index 237db181f54..97178c71b44 100644
--- a/src/plugins/task/affinity/dist_tasks.c
+++ b/src/plugins/task/affinity/dist_tasks.c
@@ -351,8 +351,7 @@ void lllp_distribution(launch_tasks_request_msg_t *req, uint32_t node_id)
}
- /* FIXME: think about core_bitmap for select/linear (all '1')
- * I'm also worried about core_bitmap with CPU_BIND_TO_SOCKETS &
+ /* FIXME: I'm worried about core_bitmap with CPU_BIND_TO_SOCKETS &
* max_cores - does select/cons_res plugin allocate whole
* socket??? Maybe not. Check srun man page.
*/
@@ -372,6 +371,21 @@ void lllp_distribution(launch_tasks_request_msg_t *req, uint32_t node_id)
#endif
/* convert masks into cpu_bind mask string */
_lllp_generate_cpu_bind(req, maxtasks, masks);
+ } else {
+ char *avail_mask = _alloc_mask(req,
+ &whole_nodes, &whole_sockets,
+ &whole_cores, &whole_threads,
+ &part_sockets, &part_cores);
+ if (avail_mask) {
+ xfree(req->cpu_bind);
+ req->cpu_bind = avail_mask;
+ req->cpu_bind_type &= (~bind_mode);
+ req->cpu_bind_type |= CPU_BIND_MASK;
+ }
+ slurm_sprint_cpu_bind_type(buf_type, req->cpu_bind_type);
+ error("lllp_distribution jobid [%u] overriding binding: %s",
+ req->job_id, buf_type);
+ error("Verify socket/core/thread counts in configuration");
}
_lllp_free_masks(maxtasks, masks);
}
--
GitLab