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Outdated suggestions cannot be applied.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.
This PR makes a number of changes to the NRT API.
Memory allocations are no longer performed through the global
MemSys
object. It turns out due to technical limitations we can't call function pointers from a cuda module they weren't compiled in, so we can't have a single globalMemSys
holding malloc and free wrappers that we can safely access from everywhere the NRT API is used. Instead, the NRT APIs will just callmalloc
andfree
directly, rather thanTheMSys.malloc(...)
.The remainder of the
MemSys
object (which amounts to the stats counting component) has been changed to be stored as a pointer which can be set by an external function. The caller will have to separately allocateTheMSys
and call the setter function before using any NRT APIs that useTheMSys
if stats are to be recorded.This PR adds the set of output functions that are needed to post-process a meminfo-tracked string array into a cudf string column including appropriately dereferencing output strings using the NRT API.