[x] add MPI_COMM_CREATE_FROM_GROUP to line 17 at beginning of 6.4.2, also MPI_INTERCOMM_CREATE_FROM_GROUPS
[x] in section describing MPI_COMM_CREATE_FROM_GROUP, maybe need to add something more about the stringtag. Maybe add some more of the text from MPI_Comm_create_group about the tag/comm arguments.
[x] fix up text in the advice to users - there's some funny font business.
[x] reference for what reverse domain name notation and example
[x] Add som verbiage for MPI_INTERCOMM_CREATE about needing to create an intracommunicator possibly not only for spawn or join, add case of Session model too.
[x] do we need verbiage about the stringtag for MPI_INTER_COMM_CREATE_FROM_GROUPS? around page 286, how unique does this need to be?
[x] MPI_GET_VERSION - Rolf says it can at any time rather than what's currently there.Around page 364, in chapter8. There may be other places for similar functions where we should check the wording.
[x] back to attributes, do we need MPI_WTIME_IS_GLOBAL back? Rolf argues we should keep this. MPI_IO? Well we think that process sets can handle this. We'll instead have a qualifier sentence before the list of MPI_TAG_UB,...MPI_WTIME_IS_GLOBAL
[x] sentence on the same page, say all communicators created by...
[x] Rolf asks about this and MPI_Comm_spawn. We're leaving that till a later date.
hppritcha
commented
2 years ago
stringtag. Maybe add some more of the text fromMPI_Comm_create_groupabout the tag/comm arguments.