-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathnotes.txt
More file actions
151 lines (107 loc) · 4.73 KB
/
Copy pathnotes.txt
File metadata and controls
151 lines (107 loc) · 4.73 KB
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
for testing, remember to set DELTA in test_command.sh
run on 10 different nodes:
okeanos-login3 /home/balawend> srun -n 10 --ntasks-per-node=1 --time=00:00:15 --account=g101-2284 hostname
nid00124
nid00126
nid00122
nid00127
nid00119
nid00123
nid00125
nid00118
nid00120
nid00121
okeanos-login3 /home/balawend> scontrol show partition okeanos
PartitionName=okeanos
AllowGroups=ALL AllowAccounts=ALL AllowQos=ALL
AllocNodes=ALL Default=YES QoS=N/A
DefaultTime=02:00:00 DisableRootJobs=NO ExclusiveUser=NO GraceTime=0 Hidden=NO
MaxNodes=UNLIMITED MaxTime=14-00:00:00 MinNodes=0 LLN=NO MaxCPUsPerNode=UNLIMITED
Nodes=nid0[0008-0062,0072-0127,0136-0190,0192-0383,0392-0446,0456-0511,0520-0574,0576-0767,0776-0831,0840-0895,0900-1151]
PriorityJobFactor=10 PriorityTier=10 RootOnly=NO ReqResv=NO OverSubscribe=EXCLUSIVE
OverTimeLimit=NONE PreemptMode=OFF
State=UP TotalCPUs=51840 TotalNodes=1080 SelectTypeParameters=NONE
JobDefaults=(null)
DefMemPerNode=UNLIMITED MaxMemPerNode=UNLIMITED
okeanos-login3 /home/balawend> sacctmgr show qos format=Name,MaxNodes,MaxJobs,MaxWall
Name MaxNodes MaxJobs MaxWall
---------- -------- ------- -----------
normal
hpc 2-00:00:00
internal 2-00:00:00
meteo 2-00:00:00
long 7-00:00:00
ReservationName=mimuww StartTime=2025-06-11T16:00:00 EndTime=2025-06-11T18:00:00 Duration=02:00:00
Nodes=nid000[50-55] NodeCnt=6 CoreCnt=144 Features=(null) PartitionName=(null) Flags=WEEKLY,SPEC_NODES
TRES=cpu=288
Users= Groups=(null) Accounts=g101-2284 Licenses=(null) State=INACTIVE BurstBuffer=(null) Watts=n/a
MaxStartDelay=(null)
ReservationName=mimuwp StartTime=2025-06-13T16:00:00 EndTime=2025-06-13T18:00:00 Duration=02:00:00
Nodes=nid000[50-55] NodeCnt=6 CoreCnt=144 Features=(null) PartitionName=(null) Flags=WEEKLY,SPEC_NODES
TRES=cpu=288
Users= Groups=(null) Accounts=g101-2284 Licenses=(null) State=INACTIVE BurstBuffer=(null) Watts=n/a
MaxStartDelay=(null)
Odpalanie:
[balawend@hpc ~/distributed-sssp]$ make test
[balawend@hpc ~/distributed-sssp]$ ssh okeanos
okeanos-login3 /home/balawend> cd testing_env/
okeanos-login3 balawend/testing_env> sbatch sbatch_run_tests.sh
okeanos-login3 balawend/testing_env> squeue | grep balawend
okeanos-login3 balawend/testing_env> cat output.txt
okeanos-login3 balawend/testing_env> srun -n 10 ./test_command.sh LOGIN69 path_20_4
# max nodes is 90!
check disk quota:
lfs quota -uh $USER /lu/tetyda/home/
zmienna: setenv TESTMAX 30000
NOTE: duze testy mi dają OOM, bo puszczam zbyt duzo procesow na jednym nodzie!
odpalanie: setenv TESTMAX 100000 ; make test-big
1c5a12744f98f331304966fe6d82eb0505e8cde9 - ten commit ma wpisana moj opt, ale wczesniejszy przechodzi na pewno
spoko testy
64567e75a1b44aa146b195e97ac20022915f3e8a - ten commit wygląda sensownie dla bigcycle
7d86ceda9b3d74f7dbdc9c5024dd5fbbe0726605 - ten commit przechodzi wszystkie testy
ZLICZANIE:
- number of phases, dla roznych DELTA, dla skal 28-30
- number of relaxations, dla roznych DELTA, skale 28-30
failing: graph500-scale-5_32_14
Running: graph500-scale-5_32_11
Running: graph500-scale-5_32_13
bash -c 'for i in $(seq 0 13); do diff "outputs/${i}.out" "tests/graph500-scale-5_32_14/${i}.out"; done'
MAM KURWA DEADLOCK Z PIZDY JAKIS
bash -c 'for f in testing_env/debug_log_*.txt; do echo -n "$f"; tail -n 1 "$f"; done' | sort | grep 'epoch: 2' | wc -l
768
TESTY wygenerowane:
RMAT1: ef16, scale 10-29 incl.
17 jun 01:10
srun -n 24 ./test_command.sh LOGIN69 random-scale20-ef16-s42_1048576_24
po 12 minutach nie sterminowal skurwysyn!
> srun -n 24 ./test_command.sh LOGIN69 random-scale16-ef16-s42_65536_24
srun: job 1393620 queued and waiting for resources
srun: job 1393620 has been allocated resources
Parsing data took: 0.068995s
Delta-stepping (one-sided) finished.
Time: 6.32449s.
Short relaxations: 23031
from which bypassed: 0
Long relaxations: 1024182
Total phases: 3144
> srun -n 48 ./test_command.sh LOGIN69 random-scale17-ef16-s42_131072_48
srun: job 1393622 queued and waiting for resources
srun: job 1393622 has been allocated resources
Parsing data took: 0.117667s
Delta-stepping (one-sided) finished.
Time: 5.11534s.
Short relaxations: 46269
from which bypassed: 0
Long relaxations: 2052865
Total phases: 7056
PO DODANIU SELF-RELAX:
> srun -n 24 --time 10:00 --tasks-per-node 24 --cpus-per-task 1 --nodes 1 -- ./test_command.sh LOGIN69 random-scale16-ef16-s42_65536_24/
srun: job 1393664 queued and waiting for resources
srun: job 1393664 has been allocated resources
Parsing data took: 0.0719039s
Delta-stepping (one-sided) finished.
Time: 6.14703s.
Short relaxations: 23031
from which bypassed: 0
Long relaxations: 1024182
Total phases: 3144