[Mpi3-rma] Chapter Reviews for One Sided Chapter
dinan at mcs.anl.gov
Sun Jul 15 22:10:40 CDT 2012
Sorry for taking so long to send my review. I've attached a marked-up
pdf with lots of edits. Most of these should be ticket 0, but I feel
that all of them are important enough to resolve somehow for MPI 3.0.
You will need real (free) adobe acrobat to view the edits (evince, for
example, will not work at all). In acrobat, select view->navigation
panels->comments to show a list of all edits and notes. Don't just scan
the document to find them or you will miss some of the small edits. I
also attached notes to explain some edits and another pdf viewer may not
be able to show them to you.
If pdf markup isn't going to work, let me know and I'll type up these
edits up as text.
One set of changes that may need some discussion are the flush
operations edits. These operations are currently described as local,
which contradicts the progress semantics for synchronization operations
in 11.7.3 (highlighted in the pdf). If you have to wait for another
process to give up the lock, a flush can become nonlocal as described in
this section. I think the best option is to classify flushes as
synchronization operations and remove the offending sentences from the
flush text since it's already covered in 11.7.3.
On 7/12/12 1:50 PM, Torsten Hoefler wrote:
> Review reminder! The meeting is next Monday!
> Thanks & Best,
> On Thu, Jul 05, 2012 at 12:40:51PM -0500, Torsten Hoefler wrote:
>> Ticket #283 is now also committed (r1244). Please review it or include
>> it in your chapter committee review!
>> I put a copy of the document for your convenience at
>> http://www.unixer.de/sec/mpi-report-r1244.pdf .
>> On Wed, Jul 04, 2012 at 05:29:14PM -0500, Torsten Hoefler wrote:
>>> [This message got stuck in mailman because the chapter committee is too
>>> large (max recipients exceeded :-) -- here is a copy for the working
>>> group -- Thanks, Torsten]
>>> Dear One Sided Chapter Committee and Working Group,
>>> Nearly all commits to the One Sided chapter for MPI-3.0 should be done
>>> by now (most of them since quite a while). Ticket #283 is unclear, so I
>>> didn't commit anything yet. Jim, can you elaborate?
>>> Please review the correct implementation of the following tickets *and*
>>> sign off *at each ticket* that you reviewed it before the next Forum (in
>>> two weeks):
>>> #140, #270, #284, #298, #308, #309
>>> Everybody is welcome to review, there can never be enough eyes checking
>>> the implementation.
>>> You can either pull a copy from subversion or use a build of r1243 that
>>> I put at: http://www.unixer.de/sec/mpi-report-r1243.pdf .
>>> Please let me know if you have any questions!
>>> ### qreharg rug ebs fv crryF ------------- http://www.unixer.de/ -----
>>> Torsten Hoefler | Performance Modeling and Simulation Lead
>>> Blue Waters Directorate | University of Illinois (UIUC)
>>> 1205 W Clark Street | Urbana, IL, 61801
>>> NCSA Building | +01 (217) 244-7736
>>> mpi3-rma mailing list
>>> mpi3-rma at lists.mpi-forum.org
>> ### qreharg rug ebs fv crryF ------------- http://www.unixer.de/ -----
>> Torsten Hoefler | Performance Modeling and Simulation Lead
>> Blue Waters Directorate | University of Illinois (UIUC)
>> 1205 W Clark Street | Urbana, IL, 61801
>> NCSA Building | +01 (217) 244-7736
>> mpi3-rma mailing list
>> mpi3-rma at lists.mpi-forum.org
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 3118239 bytes
Desc: not available
More information about the mpiwg-rma