Hi
Nodes and partions are as follows. The nodes with ThreadsPerCore=2 works
as expected, all CPU slots are in use. The small ones are all but idle,
with a few stray array jobs running a task on a single core.
Submitted jobs with no constraints end up as PENDING with
Reason=Priority, even though nodes are idle or lightly loaded. Same
result with --nodelist trying to force a job onto one of the idle nodes.
'
# COMPUTE NODES
NodeName=small1 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1 RealMemory=15982
NodeName=small2 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1 RealMemory=15982
NodeName=small19 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small23 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small24 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small25 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small29 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small30 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small31 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small32 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small33 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small34 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=small35 Sockets=2 CoresPerSocket=4 ThreadsPerCore=1
RealMemory=15982
NodeName=big20 Sockets=2 CoresPerSocket=6 ThreadsPerCore=2 RealMemory=64365
NodeName=big21 Sockets=2 CoresPerSocket=6 ThreadsPerCore=2 RealMemory=64365
NodeName=big22 Sockets=2 CoresPerSocket=6 ThreadsPerCore=2 RealMemory=64365
NodeName=big26 Sockets=2 CoresPerSocket=6 ThreadsPerCore=2 RealMemory=64365
NodeName=big27 Sockets=2 CoresPerSocket=6 ThreadsPerCore=2 RealMemory=64365
NodeName=big28 Sockets=2 CoresPerSocket=6 ThreadsPerCore=2 RealMemory=64365
NodeName=bigger1 Sockets=2 CoresPerSocket=6 ThreadsPerCore=2
RealMemory=192416
PartitionName=application Nodes=big[26-28] Default=NO MaxTime=INFINITE
Priority=65533 Shared=FORCE:400 LLN=YES State=UP
PartitionName=HPC
Nodes=small[1-2,19,23-25,29-33,34-35],big[20-22],bigger1 Default=YES
MaxTime=7-0 Shared=NO LLN=YES State=UP
'
Any suggestions? Some change in 15.08 requires changes to my config?
Thanks,
Torkil
--
Torkil Svensgaard
Sysadmin
MR-Forskningssektionen, afs. 714
DRCMR, Danish Research Centre for Magnetic Resonance
Hvidovre Hospital
Kettegård Allé 30
DK-2650 Hvidovre
Denmark
Tel: +45 386 22828
E-mail: [email protected]
