Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: Re: [OMPI devel] algorithm selection in open mpi
From: Jeff Squyres (jsquyres) (jsquyres_at_[hidden])
Date: 2012-05-07 20:55:44


George will have to answer that in detail, but note that if you modify the tuned coll module source code, you can simply "make install" in ompi/mca/coll/tuned. That will re-build the coll tuned module and install it in the plugin directory. You don't even need to recompile your MPI app, since all of the coll tuned module is dynamically opened at run-time.

That's *much* faster than re-compiling/re-installing the whole of Open MPI. w00t. :-)

On May 7, 2012, at 4:24 AM, roswan ismail wrote:

> hi all..
> i already got the results from all algorithm used in open mpi for bcast. If i want to modify binomial algorithm for example, there is a simpler way to do that? or i just need to modify "ompi_coll_tuned_bcast_intra_binomial" function, then recompile and force the system to broadcast the data using a modified binomial?? is it the right way?? thanks
>
>
> Roswan Ismail,
> FSKIK,
> Universiti Pendidikan Sultan Idris,
> Tanjong Malim, Perak, Malaysia.
> iewanis_at_[hidden]
> roswan_at_[hidden]
>
> From: George Bosilca <bosilca_at_[hidden]>
> To: Open MPI Developers <devel_at_[hidden]>
> Sent: Tuesday, April 3, 2012 9:06 PM
> Subject: Re: [OMPI devel] algorithm selection in open mpi
>
> Of course !!! ;)
>
> First set coll_tuned_use_dynamic_rules to 1, and then use coll_tuned_dynamic_rules_filename to specify a file containing the selection logic. This is kind of tricky to write, so we don't advertise it to widely. I added an example below, contact me privately if you need more info.
>
> Thanks,
> george.
>
>
> 1 # num of collectives
> 3 # ID = 3 Alltoall collective (ID in coll_tuned.h)
> 1 # number of com sizes
> 64 # comm size 8
> 2 # number of msg sizes
> 0 3 0 0 # for message size 0, bruck 1, topo 0, 0 segmentation
> 8192 2 0 0 # 8k+, pairwise 2, no topo or segmentation
> # end of first collective
>
>
> On Apr 3, 2012, at 09:01 , Pavel Mezentsev wrote:
>
>> Is there a way to specify collective depending on the size of the message and number of processes?
>>
>> Regards,
>> Pavel Mezentsev
>>
>> 2012/4/3 George Bosilca <bosilca_at_[hidden]>
>> Roswan,
>>
>> There a re simpler solutions to achieve this. We have a built-in mechanism to select a specific collective implementation. Here is what you have to add in your .openmpi/mca-params.conf (or as MCA argument on the command line):
>>
>> coll_tuned_use_dynamic_rules = 1
>> coll_tuned_bcast_algorithm = 6
>>
>> The first one activate the dynamic selection of collective algorithms, while the second one force all broadcast to be of the type 6 (binomial tree). Btw, once you set the first one, do a quick "ompi_info --param coll tuned" to see the list of all possible options for the collective algorithm selection.
>>
>> george.
>>
>> On Apr 2, 2012, at 23:10 , roswan ismail wrote:
>>
>>> Hi all..
>>>
>>> I am Roswan Ismail from Malaysia. I am focusing on MPI communication performance on quad-core cluster at my university. I used Open MPI-1.4.3 and measurements were done using scampi benchmark.
>>>
>>> As I know, open MPI used multiple algorithms to broadcast data (MPI_BCAST) such as binomial, pipeline, binary tree, basic linear and split binary tree. All these algorithms will be used based on message size and communicator size. For example, binomial is used when message size to be broadcasted is small while pipeline used for broadcasting a large message.
>>>
>>> What I want to do now is, to use fixed algorithm i.e binomial for all message size. I want to see and compare the results with the default results. So, I was modified coll_tuned_decision_fixed.c which is located in open mpi-1.4.3/ompi/mca/coll/tuned by returning binomial algorithm for all condition. Then I recompile the files but the problem is, the results obtained is same as default. It seems I do not do any changes to the codes.
>>>
>>> So could you guys tell me the right way to do that.
>>>
>>> Many thanks
>>>
>>> Roswan Binti Ismail,
>>> FTMK,
>>> Univ. Pend. Sultan Idris,
>>> Tg Malim, Perak.
>>> Pej: 05-4505173
>>> H/P: 0123588047
>>> iewanis_at_[hidden]
>>> roswan_at_[hidden]
>>> _______________________________________________
>>> devel mailing list
>>> devel_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>>
>>
>> _______________________________________________
>> devel mailing list
>> devel_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>>
>> _______________________________________________
>> devel mailing list
>> devel_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>
>
> _______________________________________________
> devel mailing list
> devel_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>
> _______________________________________________
> devel mailing list
> devel_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/devel

-- 
Jeff Squyres
jsquyres_at_[hidden]
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/