Hi all, I have updated the chapter PDF based on my notes from the last meeting I attended (10/2). The basic change was to make it clear what the contents of desc and desc_len should be in the _get_info routines for control and performance variables across connected processes. I left out changes that we discussed regarding maximum string lengths, because from the email list it looks like you all decided that that issue should be a separate ticket (with which I whole-heartedly agree). So, please have a look at it if you can before our meeting on Thursday. I know you're busy with SC preparations, but it's only two small changes! You have time! :-) Kathryn _________________________________________________________________ Kathryn Mohror, [email protected], http://scalability.llnl.gov/ Scalability Team @ Lawrence Livermore National Laboratory, Livermore, CA, USA On 8/31/14, 3:07 PM, "Schulz Martin" <[email protected]> wrote:
Hi Kathryn,
The text misses the INOUT/OUT instead of just OUT on page 34, line 10, otherwise this looks good to me.
Thanks!
Martin
________________________________________________________________________ Martin Schulz, [email protected], http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA
On 8/30/14, 4:44 PM, "Kathryn Mohror" <[email protected]> wrote:
Hi Martin,
Thanks for the feedback. I incorporated it and attached the new copy.
Anyone else have comments?
Thanks! Kathryn _________________________________________________________________ Kathryn Mohror, [email protected], http://scalability.llnl.gov/ Scalability Team @ Lawrence Livermore National Laboratory, Livermore, CA, USA
On 8/30/14, 1:22 AM, "Schulz Martin" <[email protected]> wrote:
Hi Kathryn, all,
Thanks for putting this together. I think, this looks almost right. Just the string returns are not quite right, I think. What about adding the following:
Page 15:
If the name of a control variable is the same across connected processes, the following OUT parameters must be identical: verbosity, datatype, enumtype, desc, desc_len, bind, scope. -> If the name of a control variable is the same across connected processes, the following INOUT/OUT parameters must be identical: verbosity, datatype, enumtype, desc, desc_len, bind, scope (assuming identical values are passed in desc_len).
Same on page 24: If the name and class of a performance variable are the same across connected processes, the following OUT parameters must be identical: verbosity, varclass, datatype, enumtype, desc, desc_len, bind, readonly, continuous, atomic. ->
If the name and class of a performance variable are the same across connected processes, the following INOUT/OUT parameters must be identical: verbosity, varclass, datatype, enumtype, desc, desc_len, bind, readonly, continuous, atomic (assuming identical values are passed in desc_len).
Same on page 34
If the name of a category is the same across connected processes, the following OUT parameters must be identical: desc, desc_len. -> If the name of a category is the same across connected processes, the following INOUT/OUT parameters must be identical: desc, desc_len (assuming identical values are passed in desc_len).
Martin
________________________________________________________________________ Martin Schulz, [email protected], http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA
On 8/29/14, 6:10 PM, "Kathryn Mohror" <[email protected]> wrote:
Hi all,
Here¹s a draft of the tools chapter with changes for ticket #383 based on our last discussion. Let¹s see if we can get this nailed down to read in Japan.
Please see the attached, and keep in mind that the due date for ticket announcements is MONDAY!
Thanks, Kathryn
_________________________________________________________________ Kathryn Mohror, [email protected], http://scalability.llnl.gov/ Scalability Team @ Lawrence Livermore National Laboratory, Livermore, CA, USA
_______________________________________________ mpiwg-tools mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-tools
_______________________________________________ mpiwg-tools mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-tools