Re: Running with more that 1/2 CPU's -- June 24, 2013 - 14:44

Hi Edward.

Thanks for your reply.

Allright, I can run many instances, if I have enough RAM, which I think I have. :-)

But what is the limit of one instance then?

Say I want to analyse the same dataset 100 times, each of the dataset have experienced

an external parameter change. A normal analysis create 4/5 different pipes for each 

fitting model.

Would relax be able to handle 500 pipes? 

Here I worry about the final_state file, saving, and memory management.

Is the memory released meanwhile?

Or should I flush/delete the pipes/memory for each 1/100 run?

Best

Troels

Troels Emtekær Linnet

2013/6/24 Edward d'Auvergne <edward@xxxxxxxxxxxxx>

Hi,

The only limit is the amount of RAM in your system.  I have no idea
how much memory one relax instance running a relaxation dispersion
analysis would take up.  But if you have enough you could run 10
instances of relax.  If these datasets are not part of the same
analysis then you would need to run relax 10 times separately anyway,
so in such a case the multi-processor framework from Gary Thompson
will not be of use.  The multi-processor framework would massively
speed up the process if you could only run one instance of relax at a
time.  I hope this answers your question.

Regards,

Edward

On 24 June 2013 13:24, Troels Emtekær Linnet <tlinnet@xxxxxxxxx> wrote:
> Is there a limit to how many program instances you can require of relax?
>
> If you have for example 10 different datasets you want to be analysed,
> would a work-around be to use calls to a lot of relax programs ?
>
> Is there any shared memory / temporary directories which prevent this?
>
> Best
> Troels Emtekær Linnet
>
>
> 2013/6/7 Edward d'Auvergne <edward@xxxxxxxxxxxxx>
>>
>> Hi,
>>
>> If you have OpenMPI and mpi4py installed, then you have access to Gary
>> Thompson's multi-processor framework for MPI parallelisation.  However
>> the code in relax must be written to support this.  This is the case
>> for the model-free analysis, in which case Gary has achieved near
>> perfect scaling efficiency:
>>
>> https://mail.gna.org/public/relax-devel/2007-05/msg00000.html
>>
>> However for the relaxation dispersion branch, no parallelisation has
>> been attempted, neither in the original code from Sebastian Morin or
>> the recent modifications by myself.  This is not a simple task and
>> will take a lot of effort to implement.  If this is to be implemented
>> one day, I would suggest parallelising at the level of the spin
>> clusters.
>>
>> Also note that it is often quite hard to achieve good scaling
>> efficiency and often the first attempts will just make the code
>> slower, even on a 1024 node cluster, due to the bottleneck of data
>> transfer between the nodes.  The parallelisation will also require 10
>> times as much code to be written to do the same thing as
>> non-parallised code, and debugging is much more difficult.  So while
>> it is an option, I wouldn't hold your breath for this feature.  Of
>> course unless you have the courage to tackle such a problem yourself.
>>
>> Regards,
>>
>> Edward
>>
>>
>>
>>
>> On 7 June 2013 12:24, Troels Emtekær Linnet <tlinnet@xxxxxxxxx> wrote:
>> > Hi.
>> >
>> > I have access to a computer with 24 CPU's, and want to speed things up.
>> >
>> > Architecture:          x86_64
>> > CPU op-mode(s):        32-bit, 64-bit
>> > Byte Order:            Little Endian
>> > CPU(s):                24
>> > On-line CPU(s) list:   0-23
>> > Thread(s) per core:    2
>> > Core(s) per socket:    6
>> > Socket(s):             2
>> > NUMA node(s):          2
>> > Vendor ID:             GenuineIntel
>> > CPU family:            6
>> > Model:                 44
>> > Stepping:              2
>> > CPU MHz:               2394.136
>> > BogoMIPS:              4787.82
>> > Virtualization:        VT-x
>> > L1d cache:             32K
>> > L1i cache:             32K
>> > L2 cache:              256K
>> > L3 cache:              12288K
>> > NUMA node0 CPU(s):     0,2,4,6,8,10,12,14,16,18,20,22
>> > NUMA node1 CPU(s):     1,3,5,7,9,11,13,15,17,19,21,23
>> >
>> > I have been looking at:
>> > http://www.nmr-relax.com/manual/Usage_multi_processor.html
>> >
>> > This says, that you need mpi installed.
>> >
>> > Is it possible to run relax_disp in GUI mode with more than 1/2
>> > processors?
>> >
>> > Best
>> >
>> >
>> > Troels Emtekær Linnet
>> >
>> > _______________________________________________
>> > relax (http://www.nmr-relax.com)
>> >
>> > This is the relax-users mailing list
>> > relax-users@xxxxxxx
>> >
>> > To unsubscribe from this list, get a password
>> > reminder, or change your subscription options,
>> > visit the list information page at
>> > https://mail.gna.org/listinfo/relax-users
>
>