Martin, all, I went through the issues grouped under the RMA chapter (Chapter 12) and found that some of them are issues in the I/O chapter: rows 228-259 (except for row 239) seem relevant for the I/O chapter and not the RMA chapter. I guess there has been some confusion about document and PDF numbers. I don't want to mess with the excel sheet but wanted to raise this here so that the I/O CC can take a look at these points. Cheers Joseph On 9/17/23 10:30, Martin Schulz via mpi-forum wrote:

Hi all,

First of all, thank all for a very productive and focused meeting last week on our path to MPI 4.1 – both to those attending in person in Bristol and to those online at often quite unfortunate hours!

As discussed, I compiled a list of all open points and consolidated this into a single spreadsheet, which is here:

https://docs.google.com/spreadsheets/d/1ZpsyVTnfNHv85K9Mzi_3niNXdmQ67yV2iwSa...

There is a total of 380 points that we collected that we think should be still addressed somehow. I placed them into five categories:

* MPI 4.1: This is more or less the default category and should at the end contain all small changes that we still want to apply the standard (and which will be voted on). For all items in these categories, we will need Pull requests and Issue created. * Editor changes: Small changes and formatting changes that Bill will apply directly – no separate PR or Issue is necessary. Once an item is in this category, please do not touch it (unless Bill pushes it back to one of the other categories) * Flow: Items covering bad page breaks, strange spacing, etc., which do not change the contents of the document and which need to be addressed at the very end during layout – Bill will go over them at the end and these items are mainly meant as hints for him to pay attention to * MPI 4.2: Items we think are too big and which should be pushed to the next version of the standard or separate erratas. For those, we will eventually also need PRs and Issues and we could technically start reading them in the next meeting if there is time. * Reject: items we decided to drop

Note: this list all points individually as we recorded them – multiple of them can of course (and should) be grouped into a single PR or Issue.

ACTION ITEMS:

* If you reviewed parts of the standard, please check the list and make sure comments are properly recorded (if you noticed mistakes in the page numbers, please let me know, as this means updates in hidden parts of the spreadsheet) * CC chairs: please take a look at all comments for your chapter and ensure that they are consolidated correctly. By default you are responsible for completing or shepherding all items in your chapter. * If you have created PRs and Issues (or if you do so from now on), please place them into the appropriate columns, so we can track it and place yourself into the owner category. * If the PR/Issue is complete and ready for review, please mark this in the column “DONE”. * If you think an MPI 4.1 issue should be moved to one of the other categories, please move the checkmark and note the reason for the decision in the “Comments” column at the end (together with your name).

As Wes noted on Friday, we have a short time line here – all items we want for 4.1 must be in the DONE column by September 26.

I hope this all makes sense and covers all steps forward – if you have any questions, comments or concerns, please let me know.

Thanks!

Martin

PS: This is a shared spreadsheet, so please be careful when editing and please do not change anything structural. Also, mark any decisions, observations or comments in the last column together with your name, so we have a track record for the discussions next time. Thanks!

--

Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems

Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching

Member of the Board of Directors at the Leibniz Supercomputing Centre (LRZ)

Email: [email protected]

_______________________________________________ mpi-forum mailing list [email protected] https://lists.mpi-forum.org/mailman/listinfo/mpi-forum