422 clear setelement and usage of pinnedvector leads to segementation fault#45

Merged

anyzelman merged 13 commits intodevelopfrom

422-clear-setelement-and-usage-of-pinnedvector-leads-to-segementation-fault

Aug 10, 2022

Commits on Aug 10, 2022

Better debug tracing and one super small perf. enhancement (do not call assignAll if already dense. While assignAll does the same check, this will save the call into that function altogether-- whil…
anyzelman
committed
Bugfix in grb::clear (BSP1D) detected by the new test introduced for clearing issue #422
anyzelman
committed
Issue #422: add another cleared test for good measure
anyzelman
committed
Issue #422: Now tests both SEQUENTIAL and PARALLEL modes of the pinnedVector
anyzelman
committed
Unrelated minor code style fixes
anyzelman
committed
When testing PARALLEL I/O but launching from a sequential process, then of course we only get one part of the vector for P>1. If we want to perform a test alike the one now removed, then a new nati…
anyzelman
committed
Add a new test on vectors with two entries only
anyzelman
committed
Code robustness: BSP1D PinnedVector constructor now throws on errors. Also in PARALLEL mode, if the nnz count is not updated, do that before pinning.
anyzelman
committed
The test now does size checking the moment the PinnedVector is created in the ALP/GraphBLAS context itself. It also does number-of-nonzeroes checking within the ALP/GraphBLAS context when in parall…
anyzelman
committed