1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
|
#!/usr/bin/env expect
############################################################################
# Purpose: Test of Slurm functionality
# Test of resource allocation layout. Specifically make sure that
# no excess CPUs are allocate to the job. See bug 6274.
############################################################################
# Copyright (C) 2019 SchedMD LLC
# Written by Morris Jette
#
# This file is part of Slurm, a resource management program.
# For details, see <https://slurm.schedmd.com/>.
# Please also read the included file: DISCLAIMER.
#
# 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.
############################################################################
source ./globals
set file_in "$test_dir/input"
set file_out "$test_dir/output"
set job_name $test_name
set job_id 0
proc cleanup {} {
global job_id
cancel_job $job_id
}
proc test_alloc_size { task_cnt } {
global bin_cat bin_rm file_in file_out sbatch job_name number
global alloc_unit_num max_tasks job_id
exec $bin_rm -f $file_out
set job_id [submit_job -fail "-t1 -o $file_out -J $job_name -n $task_cnt $file_in"]
if {[wait_for_job -timeout 10 $job_id "DONE"] != 0} {
log_warn "Job $job_id did not complete"
cancel_job $job_id
set max_tasks $task_cnt
return
}
wait_for_file -timeout 10 -fail $file_out
set num_cpus -1
set num_nodes -1
set threads_per_core 1
set cores_per_socket 1
set sockets_per_node 1
spawn $bin_cat $file_out
expect {
-re "NumCPUs=($number)" {
set num_cpus $expect_out(1,string)
exp_continue
}
-re "SLURM_NNODES=($number)" {
set num_nodes $expect_out(1,string)
exp_continue
}
-re "ThreadsPerCore=($number)" {
if {$threads_per_core < $expect_out(1,string)} {
set threads_per_core $expect_out(1,string)
}
exp_continue
}
eof {
wait
}
}
log_user 0
spawn $bin_cat $file_out
expect {
-re "CoresPerSocket=($number)" {
if {$cores_per_socket < $expect_out(1,string)} {
set cores_per_socket $expect_out(1,string)
}
exp_continue
}
eof {
wait
}
}
spawn $bin_cat $file_out
expect {
-re "Sockets=($number)" {
if {$sockets_per_node < $expect_out(1,string)} {
set sockets_per_node $expect_out(1,string)
}
exp_continue
}
eof {
wait
}
}
log_user 1
# Determine largest allocation unit
if {$alloc_unit_num == 4} {
set alloc_unit 1
} elseif {$alloc_unit_num == 3} {
set alloc_unit $threads_per_core
} elseif {$alloc_unit_num == 2} {
set alloc_unit [expr $threads_per_core * $cores_per_socket]
} elseif {$alloc_unit_num == 1} {
set alloc_unit [expr $threads_per_core * $cores_per_socket * $sockets_per_node]
} else {
fail "Invalid allocation unit: $alloc_unit_num"
}
set max_alloc [expr $task_cnt + (($alloc_unit - 1) * $num_nodes)]
if {$num_cpus > $max_alloc} {
log_debug "TASKS:$task_cnt"
log_debug "CPUS:$num_cpus"
log_debug "NODES:$num_nodes"
log_debug "MAX_THREADS_PER_CORE:$threads_per_core"
log_debug "MAX_CORES_PER_SOCKET:$cores_per_socket"
log_debug "MAX_SOCKETS_PER_NODE:$sockets_per_node"
log_debug "ALLOCATION_UNIT:$alloc_unit"
fail "Job with $task_cnt tasks allocated too many CPUs ($num_cpus > $max_alloc)"
}
}
if {[param_contains [get_config_param "SelectTypeParameters"] "CR_ONE_TASK_PER_CORE"]} {
skip "This test is incompatible SelectTypeParameters=CR_ONE_TASK_PER_CORE"
}
set select_type_param [get_select_type_params]
if {[check_config_select "linear"]} {
set alloc_unit_str "NODE"
set alloc_unit_num 1
} elseif { [default_part_exclusive] == 1} {
set alloc_unit_str "NODE"
set alloc_unit_num 1
} elseif { [string first "CR_SOCKET" $select_type_param] != -1} {
set alloc_unit_str "SOCKET"
set alloc_unit_num 2
} elseif { [string first "CR_CORE" $select_type_param] != -1} {
set alloc_unit_str "CORE"
set alloc_unit_num 3
} else {
set alloc_unit_str "CPU"
set alloc_unit_num 4
}
set max_tasks [get_total_cpus]
log_debug "Resource allocation unit: $alloc_unit_str"
log_debug "CPUs in default partition: $max_tasks"
if {$max_tasks > 32} {
set max_tasks 20
}
make_bash_script $file_in "
$scontrol -dd show job \$SLURM_JOB_ID | grep NumCPUs=
$scontrol -dd show job \$SLURM_JOB_ID | grep CPU_IDs=
$bin_echo ''
env | grep SLURM_NNODES=
env | grep SLURM_TASKS_PER_NODE=
env | grep SLURM_JOB_CPUS_PER_NODE=
$bin_echo ''
$scontrol show node \$SLURM_JOB_NODELIST | grep -v Features | grep -v Gres | \
grep -v CPUAlloc | grep -v NodeAddr | grep -v OS | grep -v Partitions | \
grep -v CfgTRES | grep -v AllocTRES | grep -v BootTime | grep -v Watts
"
for {set inx 1} {$inx < $max_tasks} {incr inx} {
test_alloc_size $inx
}
|
