Pack/Unpack data for user-data #1224
Conversation
Davknapp
added
New feature
and pack and unpack without prefix for the size
Pack opimization
src/t8_data/t8_data_handler.hxx
Outdated
| #include <t8.h> | ||
| #include <vector> | ||
| #include <t8_data/t8_data_handler_base.hxx> | ||
| #include <algorithm> | ||
| #include <memory> | ||
| #include <type_traits> |
There was a problem hiding this comment.
Do you maybe want to sort them?
| * \param[in] comm The MPI communicator used for the operation. | ||
| */ | ||
| virtual void | ||
| pack_vector_prefix (void *buffer, const int num_bytes, int &pos, sc_MPI_Comm comm) |
There was a problem hiding this comment.
Is there no other way than using a void pointer? Couldn't this be templatized?
There was a problem hiding this comment.
No, because the buffer used by MPI is a void* buffer. We want to pass this buffer to the MPI-function and it expects a void *
| * | ||
| * \param[in, out] buffer A pointer to the buffer where the vector prefix will be packed. | ||
| * \param[in] num_bytes The number of bytes to be packed. | ||
| * \param[in] pos A reference to an integer representing the current position in the buffer. This will be updated as bytes are packed. |
There was a problem hiding this comment.
What would I need this for?
There was a problem hiding this comment.
This is used to fill the buffer for MPI communication for data structures that are non-trivial and do not lay linear in memory.
| * | ||
| * \return An integer representing the type. | ||
| */ | ||
| virtual int |
There was a problem hiding this comment.
Which integer is for which type? Is there an enum for this?
There was a problem hiding this comment.
This is currently more a place-holder than a proper implementation. An enum over all of our types that can be used by this class would be nice.
An interface for arbitrary data.
For new user-data it suffies to implement how a single element should be packed/unpacked.
The routines for vectors of user-data make use of this routine and pack/unpack the data accordingly.
Furthermore it provides a wrapper around the send/recv routine from MPI for to avoid code duplication and easy usage. This part of the interface should be extended in the future to other MPI send/recv routines.
All these boxes must be checked by the reviewers before merging the pull request:
As a reviewer please read through all the code lines and make sure that the code is fully understood, bug free, well-documented and well-structured.
General
The reviewer executed the new code features at least once and checked the results manually
The code follows the t8code coding guidelines
New source/header files are properly added to the Makefiles
The code is well documented
All function declarations, structs/classes and their members have a proper doxygen documentation
All new algorithms and data structures are sufficiently optimal in terms of memory and runtime (If this should be merged, but there is still potential for optimization, create a new issue)
Tests
Github action
The code compiles without warning in debugging and release mode, with and without MPI (this should be executed automatically in a github action)
All tests pass (in various configurations, this should be executed automatically in a github action)
If the Pull request introduces code that is not covered by the github action (for example coupling with a new library):
Scripts and Wiki
script/find_all_source_files.scpto check the indentation of these files.Licence
doc/(or already has one)