ChemWiki - User contributions [en]

Mod:Hunt Research Group/calendar

2015-06-16T11:18:30Z

Zz6114: /* Calendar */

Back to the main [https://www.ch.ic.ac.uk/wiki/index.php/Mod:Hunt_Research_Group wiki-page]
== Calendar ==

{| class="wikitable"
|-
! 1
! 2
! 3
! 4
|-
| Tricia (Done)
| Bryan Ward (Done)
| Claire Ashworth (Not Done)
| Precious Ugbomah (Not Done)
|-
| Vincent Chen (Not Done)
| Gabriel Lau (Not Done)
| Rebecca Rowe (Not Done)
| Ken Watson (Not Done)
|-
| Laila Bana (Not Done)
| Oliver Benton (Not Done)
| Christopher Sewell (Not Done)
| Chen Chen (Not Done)
|-
| Aiswarya Prabha (Not Done)
|Ziyun Zhang (Done)
|Richard Fogarty (Not Done)
|
|-
|-)
|}

<table border="1" cellpadding="5" width="1190">
<tr bgcolor="#66CCFF">
<th width="170px">Mon</th>
<th width="170px">Tues</th>
<th width="170px">Wed</th>
<th width="170px">Thur</th>
<th width="170px">Fri</th>
<th width="170px">Sat</th>
<th width="170px">Sun</th>
</tr>

<tr valign="top">
<td bgcolor="#4357CFF">'''1st June''' </td>
<td>2nd </td>
<td>3rd </td>
<td>4th </td>
<td>5th </td>
<td bgcolor="#33CC99">6th </td>
<td bgcolor="#33CC99">7th </td>
</tr>
<tr valign="top">
<td>8th </td>
<td>9th </td>
<td>10th </td>
<td>11th </td>
<td>12th </td>
<td bgcolor="#33CC99">13th </td>
<td bgcolor="#33CC99">14th </td>
</tr>
<tr valign="top">
<td>15th </td>
<td>16th </td>
<td>17th </td>
<td>18th </td>
<td>19th Bryan Away </td>
<td bgcolor="#33CC99">20th Bryan Away </td>
<td bgcolor="#33CC99">21st Bryan Away </td>
</tr>
<tr valign="top">
<td>22nd Bryan Away </td>
<td>23rd Bryan Away </td>
<td>24th Bryan Away </td>
<td>25th Bryan Away </td>
<td>26th Bryan Away </td>
<td bgcolor="#33CC99">27th Bryan Away </td>
<td bgcolor="#33CC99">28th Bryan Away </td>
</tr>
<tr valign="top">
<td >29th Bryan Away </td>
<td>30th </td>
<td bgcolor="#4357CFF">'''1st July''' </td>
<td>2nd </td>
<td>3rd Tricia Away </td>
<td bgcolor="#33CC99">4th Tricia Away </td>
<td bgcolor="#33CC99">5th Tricia Away </td>
</tr>
<tr valign="top">
<td >6th Tricia Away </td>
<td>7th </td>
<td>8th </td>
<td>9th </td>
<td>10th </td>
<td bgcolor="#33CC99">11th </td>
<td bgcolor="#33CC99">12th </td>
</tr>
<tr valign="top">
<td>13th </td>
<td>14th </td>
<td>15th </td>
<td>16th </td>
<td>17th </td>
<td bgcolor="#33CC99">18th </td>
<td bgcolor="#33CC99">19th </td>
</tr>
<tr valign="top">
<td>20th </td>
<td>21st </td>
<td>22nd </td>
<td>23rd </td>
<td>24th </td>
<td bgcolor="#33CC99">25th </td>
<td bgcolor="#33CC99">26th </td>
</tr>
<tr valign="top">
<td>27th </td>
<td>28th </td>
<td>29th Tricia Away </td>
<td>30th Tricia Away </td>
<td>31st Tricia Away </td>
<td bgcolor="#4357CFF">'''1st August''' Tricia Away </td>
<td bgcolor="#33CC99">2nd Tricia Away </td>
</tr>
<tr valign="top">
<td>3rd Tricia Away </td>
<td>4th Tricia Away </td>
<td>5th Tricia Away </td>
<td>6th Tricia Away </td>
<td>7th Tricia Away </td>
<td bgcolor="#33CC99">8th Tricia Away </td>
<td bgcolor="#33CC99">9th Tricia Away </td>
</tr>
<tr valign="top">
<td>10th Tricia Away </td>
<td>11th Tricia Away </td>
<td>12th Tricia Away </td>
<td>13th Tricia Away </td>
<td>14th Tricia Away </td>
<td bgcolor="#33CC99">15th Tricia Away </td>
<td bgcolor="#33CC99">16th Tricia Away </td>
</tr>
<tr valign="top">
<td>17th </td>
<td>18th </td>
<td>19th </td>
<td>20th </td>
<td>21st </td>
<td bgcolor="#33CC99">22nd </td>
<td bgcolor="#33CC99">23rd </td>
</tr>
<tr valign="top">
<td>24th </td>
<td>25th </td>
<td>26th </td>
<td>27th </td>
<td>28th </td>
<td bgcolor="#33CC99">29th </td>
<td bgcolor="#33CC99">30th </td>
</tr>
<tr valign="top">
<td>31st </td>
<td bgcolor="#4357CFF">'''1st September''' </td>
<td>2nd </td>
<td>3rd </td>
<td>4th </td>
<td bgcolor="#33CC99">5th </td>
<td bgcolor="#33CC99">6th </td>
</tr>
<tr valign="top">
<td>7th </td>
<td>8th </td>
<td>9th </td>
<td>10th </td>
<td>11th </td>
<td bgcolor="#33CC99">12th Tricia Away </td>
<td bgcolor="#33CC99">13th Tricia Away </td>
</tr>
<tr valign="top">
<td>14th Tricia Away </td>
<td>15th Tricia Away </td>
<td>16th Tricia Away </td>
<td>17th Tricia Away </td>
<td>18th Tricia Away </td>
<td bgcolor="#33CC99">19th Tricia Away </td>
<td bgcolor="#33CC99">20th Tricia Away </td>
</tr>
<tr valign="top">
<td>21st </td>
<td>22nd </td>
<td>23rd </td>
<td>24th </td>
<td>25th </td>
<td bgcolor="#33CC99">26th </td>
<td bgcolor="#33CC99">27th </td>
</tr>
<tr valign="top">
<td>28th </td>
<td>29th </td>
<td>30th </td>
<td bgcolor="#4357CFF">'''1st October''' </td>
<td>2nd </td>
<td bgcolor="#33CC99">3rd </td>
<td bgcolor="#33CC99">4th </td>
</tr>
</table>

Mod:Hunt Research Group/hpc

2015-06-02T16:53:29Z

Zz6114:

HPC Notes (Ling and Tricia)

Back to the main [https://www.ch.ic.ac.uk/wiki/index.php/Mod:Hunt_Research_Group wiki-page]

1. Queues to run on
:'''pqph''' (various, see below) this is the hunt group queue, runs on the servers listed below
:'''pqchem''' (42 nodes) this is the chemistry department queue,
:'''chemlab1''' (2 chassis) this is for the computational chemistry lab but can be used out of term time

:submitting jobs
::various for pqph
::pqchem and chemlab1 work best with job submitted on 8 cores

2. submit script uses PBS queing
::\#batch processing commands
::\#PBS -l walltime=119:59:00
::\#PBS -lselect=1:ncpus=8:mem=15800MB:tmpspace=400gb
::\#PBS -j oe
::\#PBS -q pqph

:alter -q for the relevant queue
:alter -l to the walltime of the relevant queue (use qstat -q to find the walltime)

:add tmpspace=400 to ensure large disk jobs get put on a node with enough disk!! Note that this requires you to include
maxdisk=400gb in your gaussian input.
:NOTE the queing system does not check disk allocations. When requesting large disk jobs remember to request all of the processors on a node even if you are not using all of the processors. For large jobs the maximum disk space you can request is '''800GB''' on the 12 processor nodes.

Recommended specification is
:in PBS script
::ncpus=8
::mem=15800MB
::tmpspace=400gb

:in gaussian file
::%nprocshared=8
::%mem=14400MB
::maxdisk=400GB

For small jobs only
:ncpus=4
:mem=8GB
:no tmpspace=400gb specification

Using qstat
:qstat to get your jobs that are running
:qstat -q to get a list of all queues
:qdel and ID number to remove a job from the queue

3. Useful link:
:https://scanweb.cc.ic.ac.uk/uportal2/

4. Servers:
:old nodes all single core
::all 18GB disk, a couple have 80GB disks
:::2 x 8 GB 2 proc 1.8GHz
:::1 x 6 GB 2 proc 1.8GHz
:::4 x 4 GB 2 proc 1.8GHz
:::1 x 3 GB 4 proc 2.2GHz
:::1 x 4GB Opteron 244 (1.8GHz) 18GB on /tmp cx1-50-1-4
:::1 x 8GB Opteron 248 (2.2GHz) 18GB on /tmp cx1-50-1-6
:::1 x 2GB Opteron 244 (1.8GHz) 18GB on /tmp cx1-50-1-9

:newer nodes
::2 x 32 GB 4 proc 2.8GHz 160GB disk
::4 x 16 GB 8 proc 2.4GHz 500GB disk
::4MB x 48 GB 12 proc 2.66 GHz 1TB disk

:::so what does all this mean
:::1 node =1 box in a stack which forms the cluster
:::with 4 motherboards (MB) (single board groups CPU and RAM and internal buses which communicate with other parts)
:::each with 2 CPUs (central processing unit)
:::each CPU has internally six cores (cores which do the actual number crunching)
:::so one MB has 12 cores
:::and one node has 4MB*12cores=48 cores
:::each MB has 48GB memory from 12cores each with 4GB
:::each MB has 1TB disk

pqchem can handle
:: 8 proc nodes with 16 or 24 or 32 or 48 GB and 450GB (max) disk
:::8 processors 48 GBytes memory 225 GByte disk (8 nodes)(nehalem)
:::8 processors 32 GBytes memory 450 GByte disk (tmp400=true) (10 nodes)(harpertown)
:::8 processors 48 GBytes memory 450 GByte disk (tmp400=true) (4 nodes) (nehalem)
:::8 processors 24 GBytes memory 450 GByte disk (tmp400=true) (20 nodes) (nehalem)

5. memory and disk requirements
::1MW is about 8.4MB
::Gaussian is greedy and will exceed the allocated memory always leave about 50MB extra per processor, eg if you have 8000MB=8GB available for 8 nodes then request in the %mem=7600MB
:: each proc needs a gaussian executable, which takes about 8mw (or 12 for MP2 frequencies)
:::so 4 proc jobs require 4*8*8.4=269MB just to run
:::so 8 proc jobs require 8*8*8.4=538MB just to run
::then normal jobs will need 2*N^2 W *8.4 to get B (1,048,576B =1MB)
::: so 300 basis functions will need 180000W =0.18MW =1.5MB in addition to the above requirements

::require 2ON^2 W of disk to run where O=number of occupied orbitals, N=number of basis functions

::MP2 jobs
:: work best with %mem and maxdisk defined
::: in-core requires N^4/4 divided by 1,000,000 MW memory
:::: so 400 basis functions will need 6400MW=53760MB=54GB memory per node, which is unlikely!
::: semi-direct requires 2*O(N^2) memory and N^3 disk
:::: so N=476 basis functions O=56 occupied orbitals will need
::::25.4MW=214MB of memory
::::and 108MW=906MB disk (this is not actually true it will need much more probably around 1800MB disk per processor!)
::::so total memory for MP2 freq 8proc will be
:::::12*8*8.4=807MB to run and 8*214=1712MB for calcs and some extra 400MB=3019MB=3.3GB
:::::gaussian does not like GB directive so give %mem in MB

6. checkpoint files
: checkpoint files should be exactly the same name as the input file name
: for jobs that may exceed the wall time there are two options
:option one: specify the full path of the checkpoint file, for example
::%chk=/work/phunt/tmp/filename.chk
:option two is to change your runscript and input file
::in the gaussian job have %chk=chk
::in the runscript replace g03 input.com with
::pbsexec g03 input.com
::mkdir $HOME/$PBS_JOBID
::cp chk $HOME/$PBS_JOBID
::this will cause Gaussian to be killed 10 minuets before the walltime limit and the checkpoint file saved out to the directory $HOME/$PBS_JOBID

7. Problems
:Without "#!/bin/sh" in the beginning of the job script, the job will always go to the queue "short" instead of the queue asked for.

8. run scripts:

:3.1 CPMD:
::https://www.ch.ic.ac.uk/wiki/index.php/Image:Runcpmd_md.sh

:3.2 DL-POLY:
::https://www.ch.ic.ac.uk/wiki/index.php/Image:Mpirun.sh

::Note: You´ll not be able to see the output until the job finishes : the directory /tmp/pb.XXX isn´t accessible to you because it is on the private disk of the node running the job.
::To get DLPOLY to terminate before the job hits the walltime limit and killed, you need to run it through a program called pbsexec, for example:
::pbsexec mpiexec DLPOLY.X
::This will kill DLPOLY 15 minutes before the walltime limit, giving your script time to transfer files back to $work.