> So it sounds like the $pe_slots=serial behavior isn't true for your grid
> engine.
> An alternative would be to convert single node jobs to serial jobs that
> requests
> a consumable(other than slots) representing the number of cores on the node.
>
> However this will change the behavior
On Fri, 31 Jul 2015 08:00:30 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> > What happens with serial jobs (ie no PE at all)? If they exhibit
> > the same behavior then we just need to figure out how to tweak the
> > scheduler config to fill up nodes. If serial jobs are already
>> complex_valuesoccupied_slots=12,slots=12,num_proc=12,h_vmem=45.01G, \
> > exclusive=true
>
> To streamline this (but it won't solve your problem): I usually recommend not
> to set num_proc,
> as it's a feature of the host which is detected by SGE already. Limitatio
> Am 31.07.2015 um 10:35 schrieb Winkler, Ursula (ursula.wink...@uni-graz.at)
> :
>
>
>> Have you got slots defined per host or just per queue?
>> It may need to be defined at a host level via complex_values
>> in order to work.
>
> Both:
>
> # qconf -sc | grep "#name\|slots"
> #name
> Have you got slots defined per host or just per queue?
> It may need to be defined at a host level via complex_values
> in order to work.
Both:
# qconf -sc | grep "#name\|slots"
#name shortcut typerelop requestable consumable
default urgency
occupied_slots o_s
On Fri, 31 Jul 2015 05:19:32 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> >> Well, I created an additional PE with alloacation_rule
> >> "$pe_slots", and built in an if condition in "pe.jsv" for all jobs
> >> which request just a single node to be assigned to this new PE.
> >> Bu
> What happens with serial jobs (ie no PE at all)? If they exhibit the same
> behavior then we
> just need to figure out how to tweak the scheduler config to fill up nodes.
> If serial jobs are
> already clumping onto a few nodes then the claim that $pe_slots behaves like
> serial on the
>
On Thu, 30 Jul 2015 12:57:13 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> > My suggestion was to modify your jsv/gepetools to force single node
> > parallel jobs into PEs with $pe_slots allocation rules (which gives
> > you control over where they are scheduled via queue_sort_met
>> Well, I created an additional PE with alloacation_rule "$pe_slots",
>> and built in an if condition in "pe.jsv" for all jobs which request
>> just a single node to be assigned to this new PE. But the annoying
>> situation didn't change. The scheduler configuration is set to
>> "queue_sort_met
> Sorry to step in the discussion: `qstat -j ...` shows the requested one, the
> granted one is in `qstat -r`.
>
> $ qsub -pe "*" 2 test.sh
> Your job 44329 ("test.sh") has been submitted $ qstat -j 44329 ...
> parallel environment: * range: 2
> ...
My jobs:
qstat -j ...
...
Am 30.07.2015 um 18:29 schrieb "Reuti" :
>
>> Am 30.07.2015 um 18:14 schrieb Winkler, Ursula (ursula.wink...@uni-graz.at)
>> :
>>
>>
>>
>>> On Thu, 30 Jul 2015 12:57:13 +
>>> "Winkler, Ursula (ursula.wink...@uni-graz.at)"
>>> wrote:
>>>
> My suggestion was to modify your jsv/gep
> Am 30.07.2015 um 18:14 schrieb Winkler, Ursula (ursula.wink...@uni-graz.at)
> :
>
>
>
>> On Thu, 30 Jul 2015 12:57:13 +
>> "Winkler, Ursula (ursula.wink...@uni-graz.at)"
>> wrote:
>>
My suggestion was to modify your jsv/gepetools to force single node
parallel jobs into PEs wi
> On Thu, 30 Jul 2015 12:57:13 +
> "Winkler, Ursula (ursula.wink...@uni-graz.at)"
> wrote:
>
>>> My suggestion was to modify your jsv/gepetools to force single node
>>> parallel jobs into PEs with $pe_slots allocation rules (which gives
>>> you control over where they are scheduled via queu
> I believe it should be a load_formula of -slots so the more slots are
> available(fewest used) the lower the load and the more attractive the
> node. The page Reuti pointed to manages to write this both ways around.
I'll try it out tomorrow - I'm not at the office now and it's a little bit
di
>> Well, I created an additional PE with alloacation_rule "$pe_slots",
>> and built in an if condition in "pe.jsv" for all jobs which request
>> just a single node to be assigned to this new PE. But the annoying
>> situation didn't change. The scheduler configuration is set to
>> "queue_sort_metho
On Thu, 30 Jul 2015 12:57:13 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> > My suggestion was to modify your jsv/gepetools to force single node
> > parallel jobs into PEs with $pe_slots allocation rules (which gives
> > you control over where they are scheduled via queue_sort_met
On Thu, 30 Jul 2015 12:57:13 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> > My suggestion was to modify your jsv/gepetools to force single node
> > parallel jobs into PEs with $pe_slots allocation rules (which gives
> > you control over where they are scheduled via queue_sort_met
On Thu, 30 Jul 2015 12:57:13 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> > My suggestion was to modify your jsv/gepetools to force single node
> > parallel jobs into PEs with $pe_slots allocation rules (which gives
> > you control over where they are scheduled via queue_sort_met
> My suggestion was to modify your jsv/gepetools to force single node
> parallel jobs into PEs with $pe_slots allocation rules (which gives
> you control over where they are scheduled via queue_sort_method and
> load_formula) while sending the others to PEs with other (appropriate)
>allocation rul
> My comment was meant to be read in the context of:
> i)Reuti's earlier suggestion that jobs in PEs with $pe_slots allocation_rule
> will be scheduled according to the queue_sort_method and load_formula like
> serial jobs
> ii)Your response that settting the allocation rule to $pe_slots stops
On Thu, 30 Jul 2015 08:25:05 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> > $pe_slots restricts you to a single node so I'm guessing the jobs
> > that don't start are jobs that need more than one node.
>
> Yes, that should be right.
>
> > While we don't use gepetools we do h
> $pe_slots restricts you to a single node so I'm guessing the jobs that don't
> start are jobs that need more than one node.
Yes, that should be right.
> While we don't use gepetools we do have a JSV that rewrites people's
> requested PE based on the number
> What you need I think is someth
...@uni-graz.at)
> Cc: users@gridengine.org
> Betreff: Re: [gridengine users] Filling up nodes when using gepetools
>
> Hi,
>
> >> Am 29.07.2015 um 12:50 schrieb Winkler, Ursula
> >> (ursula.wink...@uni-graz.at) :
> >>
> >>>> Node1 has 12 Cores/Slots
-Ursprüngliche Nachricht-
Von: Reuti [mailto:re...@staff.uni-marburg.de]
Gesendet: Mittwoch, 29. Juli 2015 15:10
An: Winkler, Ursula (ursula.wink...@uni-graz.at)
Cc: users@gridengine.org
Betreff: Re: [gridengine users] Filling up nodes when using gepetools
Hi,
>> Am 29.07.2015
Hi,
> Am 29.07.2015 um 12:50 schrieb Winkler, Ursula (ursula.wink...@uni-graz.at)
> :
>
>>> Node1 has 12 Cores/Slots and 1 MPI-Job with 2 Slots is running on it.
>>> A user submits job2 which require maximal 10 slots.
>>> Independently from schedule_interval, job_load_adjustments,
>>> load_form
>> Node1 has 12 Cores/Slots and 1 MPI-Job with 2 Slots is running on it.
>> A user submits job2 which require maximal 10 slots.
>> Independently from schedule_interval, job_load_adjustments,
>> load_formula and/or load_adjustment_decay_time parameters-settings
>> job2 usually won't start on Node1
On Wed, 29 Jul 2015 08:02:41 +
"Winkler, Ursula (ursula.wink...@uni-graz.at)"
wrote:
> Hi Gridengine users group members,
>
> I have the following problem - let me explain it with following
> example:
>
> Node1 has 12 Cores/Slots and 1 MPI-Job with 2 Slots is running on it.
> A user submits
Hi Gridengine users group members,
I have the following problem - let me explain it with following example:
Node1 has 12 Cores/Slots and 1 MPI-Job with 2 Slots is running on it.
A user submits job2 which require maximal 10 slots.
Independently from schedule_interval, job_load_adjustments, load_fo
28 matches
Mail list logo