This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision Last revision Both sides next revision | ||
cluster:119 [2013/09/24 15:28] hmeij [lava.mvampich2.wrapper] |
cluster:119 [2017/10/18 09:10] hmeij07 |
||
---|---|---|---|
Line 1: | Line 1: | ||
\\ | \\ | ||
**[[cluster: | **[[cluster: | ||
- | |||
- | Jobs need to be submitted to the scheduler on host sharptail itself for now and will be dispatched to nodes n33-n37 in queue mwgpu. | ||
- | --- // | ||
==== Submitting GPU Jobs ==== | ==== Submitting GPU Jobs ==== | ||
+ | |||
+ | Please plenty of time between multiple GPU job submissions. | ||
+ | |||
+ | Jobs need to be submitted to the scheduler on host sharptail itself for now and will be dispatched to nodes n33-n37 in queue mwgpu. They can also be submitted from host greentail but remember any output will be in shraptail' | ||
+ | --- // | ||
+ | |||
+ | **Articles** | ||
+ | |||
+ | * [[http:// | ||
+ | * [[http:// | ||
+ | |||
Line 44: | Line 52: | ||
</ | </ | ||
- | With '' | + | With '' |
< | < | ||
Line 58: | Line 66: | ||
3 Tesla K20m 21 C 0 % | 3 Tesla K20m 21 C 0 % | ||
==================================================== | ==================================================== | ||
+ | |||
+ | [hmeij@sharptail sharptail]$ ssh n33 gpu-free | ||
+ | 1,3,0 | ||
+ | |||
+ | |||
</ | </ | ||
Line 121: | Line 134: | ||
#BSUB -q mwgpu | #BSUB -q mwgpu | ||
#BSUB -J test | #BSUB -J test | ||
+ | |||
+ | # from greentail we need to set up the module env | ||
+ | export PATH=/ | ||
+ | / | ||
+ | / | ||
+ | / | ||
+ | / | ||
+ | / | ||
+ | / | ||
+ | export LD_LIBRARY_PATH=/ | ||
+ | / | ||
+ | / | ||
+ | / | ||
+ | / | ||
+ | / | ||
+ | |||
## leave sufficient time between job submissions (30-60 secs) | ## leave sufficient time between job submissions (30-60 secs) | ||
Line 158: | Line 187: | ||
# NAMD | # NAMD | ||
- | # from greentail we need to recreate module env | ||
- | export PATH=/ | ||
- | / | ||
- | / | ||
- | / | ||
- | / | ||
- | / | ||
- | / | ||
- | export LD_LIBRARY_PATH=/ | ||
- | / | ||
- | / | ||
- | / | ||
- | / | ||
- | / | ||
# signal that this is charmrun/ | # signal that this is charmrun/ | ||
export CHARMRUN=1 | export CHARMRUN=1 | ||
Line 229: | Line 244: | ||
##BSUB -q mwgpu | ##BSUB -q mwgpu | ||
##BSUB -n 1 | ##BSUB -n 1 | ||
- | ##BSUB -R " | + | ##BSUB -R " |
## signal GMXRC is a gpu run with: 1=thread_mpi | ## signal GMXRC is a gpu run with: 1=thread_mpi | ||
#export GMXRC=1 | #export GMXRC=1 | ||
Line 240: | Line 255: | ||
#BSUB -q mwgpu | #BSUB -q mwgpu | ||
#BSUB -n 1 | #BSUB -n 1 | ||
- | #BSUB -R " | + | #BSUB -R " |
# signal GMXRC is a gpu run with: 2=mvapich2 | # signal GMXRC is a gpu run with: 2=mvapich2 | ||
export GMXRC=2 | export GMXRC=2 | ||
Line 267: | Line 282: | ||
/ | / | ||
/ | / | ||
+ | export PATH=/ | ||
export LD_LIBRARY_PATH=/ | export LD_LIBRARY_PATH=/ | ||
/ | / | ||
Line 282: | Line 298: | ||
#BSUB -q mwgpu | #BSUB -q mwgpu | ||
#BSUB -n 1 | #BSUB -n 1 | ||
- | #BSUB -R " | + | #BSUB -R " |
# signal MATGPU is a gpu run | # signal MATGPU is a gpu run | ||
export MATGPU=1 | export MATGPU=1 | ||
Line 429: | Line 445: | ||
exit $? | exit $? | ||
+ | |||
+ | |||
+ | </ | ||
+ | |||
+ | |||
+ | ===== elim code ===== | ||
+ | |||
+ | < | ||
+ | |||
+ | # | ||
+ | |||
+ | while (1) { | ||
+ | |||
+ | $gpu = 0; | ||
+ | $log = ''; | ||
+ | if (-e "/ | ||
+ | $tmp = `/ | ||
+ | @tmp = split(/ | ||
+ | foreach $i (0..$#tmp) { | ||
+ | ($a, | ||
+ | if ( $f == 0 ) { $gpu = $gpu + 1; } | ||
+ | #print "$a $f $gpu\n"; | ||
+ | $log .= " | ||
+ | } | ||
+ | } | ||
+ | # nr_of_args name1 value1 | ||
+ | $string = "1 gpu $gpu"; | ||
+ | |||
+ | $h = `hostname`; chop($h); | ||
+ | $d = `date +%m/ | ||
+ | foreach $i (' | ||
+ | if ( " | ||
+ | `echo " | ||
+ | } | ||
+ | } | ||
+ | |||
+ | # you need the \n to flush -hmeij | ||
+ | # you also need the space before the line feed -hmeij | ||
+ | print " | ||
+ | # or use | ||
+ | # | ||
+ | |||
+ | # smaller than specified in lsf.shared | ||
+ | sleep 10; | ||
+ | |||
+ | } | ||