Allow parallel filters feature for comm size of 1 #840
Conversation
jhendersonHDF
requested review from
derobins,
fortnern,
jrmainzer,
qkoziol,
soumagne and
vchoi-hdfgroup
as code owners
jhendersonHDF
force-pushed
the
parallel_filters_serial_writes
branch
from
437a611 to
db3fb5c
Compare
Would this be cases like initializing with MPI_Init_thread()? EDIT: There are two other places in the library where we check for a comm_size > 1 and then do something in parallel if we're just using a comm_size of 1. One is just an optimization in H5Fsuper.c for reading the superblock from only 1 rank, so I don't think that one really matters if the communicator behaves differently from a communicator with >1 ranks. The other is in the page buffer code where we DON'T bypass the page buffer in parallel if comm_size is 1. In that case, I think this also might have the same issue? However, that code is currently #ifdef'ed out because apparently that check stopped working at some point. I think the main takeaway is, if checking for comm_size of 1 isn't good enough here, we probably shouldn't merge this or should add extra checks here and maybe in the page buffer code as well. |
* Use internal version of H5Eprint2 to avoid possible stack overflow (#661) * Add support for parallel filters to h5repack (#832) * Allow parallel filters feature for comm size of 1 (#840) * Avoid popping API context when one wasn't pushed (#848) * Fix several warnings (#720) * Don't allow H5Pset(get)_all_coll_metadata_ops for DXPLs (#1201) * Fix free list tracking and cleanup cast alignment warnings (#1288) * Fix free list tracking and cleanup cast alignment warnings * Add free list tracking code to H5FL 'arr' routines * Fix usage of several HDfprintf format specifiers after HDfprintf removal (#1324) * Use appropriate printf format specifiers for haddr_t and hsize_t types directly (#1340) * Fix H5ACmpio dirty bytes creation debugging (#1357) * Fix documentation for H5D_space_status_t enum values (#1372) * Parallel rank0 deadlock fixes (#1183) * Fix several places where rank 0 can skip past collective MPI operations on failure * Committing clang-format changes Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com> * Fix a few issues noted by LGTM (#1421) * Fix cache sanity checking code by moving functions to wider scope (#1435) * Fix metadata cache bug when resizing a pinned/protected entry (v2) (#1463) * Disable memory alloc sanity checks by default for Autotools debug builds (#1468) * Committing clang-format changes Co-authored-by: github-actions <41898282+github-actions[bot]@users.noreply.github.com>
No description provided.