[Mpi3-rma] adam's comment on chapter 11
htor at illinois.edu
Sun Jul 22 12:53:04 CDT 2012
On Thu, Jul 19, 2012 at 05:16:36PM +0000, Howard Pritchard wrote:
> Chapter 11 review things to fix:
> -----Original Message-----
> From: Kathryn Mohror [mailto:kathryn at llnl.gov]
> Sent: Thursday, July 19, 2012 9:04 AM
> To: Howard Pritchard
> Subject: adam's comment on chapter 11
> 4) The Window Info section 11.2.7 lists calls that assign info to
> windows, but it is missing the new creation routines, i.e.,
> MPI_WIN_ALLOCATE, MPI_WIN_ALLOCATE_SHARED, MPI_WIN_DYNAMIC. These
> should be included in the list that starts with "Hints are specified on
> a per window basis, in ..."
Correct, I replaced MPI_WIN_CREATE with "window creation functions" to
avoid those problems in the future.
> It would be nice if the latex formatting of the MPI_Get_accumulate
> on page 430.
> It would be nice if the latex formatting of the MPI_Rget_accumulate
> on page 437.
Wow, both of those issues could be due to the long argument lists. Bill,
do you know a fix or do we need a funcdef3 environment?
> Funny formatting also occurring on page 456 lines 13,14 for the table
> of error classes.
Hmm, this seems to be an issue with the \MPIupdate macro :-(. Bill, do
you have any idea?
> Need to review lines 15-19 on page 462 because statement about
> polling on X is not correct in the separate memory model.
They seem fine to me but I may be looking at the wrong lines (I am
working with r1501). Which version did you use?
> The example links on page 465 lines 37 and 41 aren't really correct,
> they're linked to the section numbers not the examples.
I can't find those either.
I was working off version 1501:
Thanks & Best,
### 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
More information about the mpiwg-rma