[mpich-discuss] generating slog2 format trace data

[mhsuan]

Please don't send me messages I don't want. Please exclude me when you reply
to requests of others. Thanks.

M-H

-----Original Message-----
From: mpich-discuss-bounces at mcs.anl.gov
[mailto:mpich-discuss-bounces at mcs.anl.gov] On Behalf Of Jeffrey S. Brown
Sent: Thursday, August 12, 2010 6:10 AM
To: Anthony Chan
Cc: mpich-discuss at mcs.anl.gov; jeffb at lanl.gov
Subject: Re: [mpich-discuss] generating slog2 format trace data

that should do it - thanks

Jeff

>
> Hi Jeff,
>
> We don't have a very detailed documentation on the slog2
> data format.  However, we have an API, called TRACE-API,
> that allows one to generate slog2 file programmatically
> (Several other groups had successfully used the API)
> The API is documented from p.22 to p.41 in the following document
>
> http://ftp.mcs.anl.gov/pub/mpi/slog2/slog2-draft.pdf
>
> Also, there is a sample implementation of TRACE-API
> in slog2sdk/trace_sample which should help to implement
> the callback API.
>
> A formal slog2 paper that describes the general
> features/structures of slog2 file:
>
> http://ftp.mcs.anl.gov/pub/mpi/slog2/slog2-paper.pdf
>
> Let me know if you have any questions.
>
> A.Chan
>
>
>
> ----- "Jeffrey S. Brown" <jeffb at lanl.gov> wrote:
>
>> folks,
>>
>> We are interested in hooking up a locally developed performance
>> measurement tool to jumpshot.  The idea is for our tool to generate
>> slog2 format
>> trace data then feed to jumpshot for visualization.  In order to do
>> this we need detailed documentation on the slog2 data format.  Can you
>> please
>> point us in the right direction?
>>
>> thanks,
>>
>> Jeff Brown, Los Alamos National Laboratory
>>
>> _______________________________________________
>> mpich-discuss mailing list
>> mpich-discuss at mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>

_______________________________________________
mpich-discuss mailing list
mpich-discuss at mcs.anl.gov
https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss