OMPD 5 and limited OMPD support for cuda devices #49
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit aligns libompd and gdb-wrapper with a newer OpenMP spec. * Some functionality was desiabled as it relies on information no longer exposed via OMPD. * The OMPD callback functions still have non-standard names (but signatures should be correct now) * Cuda device initialization should work now, but not much other OMPD device funtionality.
* Code adapted from Marty Mcfadden's (<[email protected]>) code for the clang-ykt runtime. * OMPD code in the nvptx device runtime library in libomptarget is currently included regardless of compile flags/defines
* libompd can retrieve thread state information for a cuda thread * ODB tries to summarize states of many cuda threads (this is still a bit ugly and doesnt repsect blocks) * Currently openmp cuda threads can only be in an undefined state or parallel work state (parallel work is not consistent)
* Moved some removed/renamed types out of ompd.h to ompd-private.h (we still need them internally i think) * Removed dead code * Some types have been renamed Still need to rename all callbacks and check for completeness
Conflicts: libompd/src/CMakeLists.txt libompd/src/omp-debug.cpp runtime/src/ompd-specific.cpp
ODB now initializes the openmp process with ompd when the first ompd commadn is issued, not when the debugger starts. This is necessary because we now check for state tracking in the process initialization function, and OpenMP runtime may not be present when the debugger has just loaded the program.
libompd does not use the new operator and instead uses the tool provides malloc callback directly.
Its not part of the spec and not used by libompd.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Aligns the OMPD interface with the draft specification of OpenMP 5 and adds limited OMPD support for cuda devices (thread handles, some thread states, parallel handels).