Ticket #7782 (closed: fixed)
System test performance monitoring: Use MemoryStats::residentMem
Reported by: | Russell Taylor | Owned by: | Russell Taylor |
---|---|---|---|
Priority: | minor | Milestone: | Release 3.0 |
Component: | Tools | Keywords: | |
Cc: | Blocked By: | ||
Blocking: | Tester: | Martyn Gigg |
Description
...so long as it gives the same or a better answer.
I didn't realise that this had been added to Mantid when I implemented #6169. Clearly it's the better thing to use so long as it works correctly.
Change History
comment:2 Changed 7 years ago by Russell Taylor
- Status changed from inprogress to verify
- Resolution set to fixed
Tester: The branch is feature/7782_use_residentmem in the systemtests repository.
If you look at the output from the system tests on Windows (see here), the memory number is much more stable after this change (which happened at 'revision' 14).
The linux number (see here) mostly got quite a bit 'worse', but really I don't think we believe it with either way of calculating it. For now, we probably should only pay attention to the numbers from Windows.
comment:3 Changed 7 years ago by Martyn Gigg
- Status changed from verify to verifying
- Tester set to Martyn Gigg
comment:4 Changed 7 years ago by Martyn Gigg
Changes look fine and agreed that we really don't quite understand the linux numbers properly yet.
Re #7782. Use residentMem instead of availMem.
It *should* give about the same answer, but without the approximation that nothing else on the system has changed.
Changeset: 5c40d718a3d95eeb9a15330931b80cded1f6936d