DavidFair
commented
2 years ago Ah looking at #47 we would need to come up with a way of loading Jemalloc with Mantid:
The package does not make sense for conda as the startup scripts are skipped.
This looks like it's blocked by upstream work so I'll close it for the moment and have a think about it :thinking:
We use Jemalloc on our build in the main repo, as this allocator will aggressively release memory. This improves the user experience when they finish processing 20GB+ data, then switch to an analysis tool (requiring that memory)
By setting this ON we bring the conda recipe in-line with upstream builds