Ticket #2568 (closed: fixed)
Analysis on windows is going to ManagedWorkspaces far too early
| Reported by: | Martyn Gigg | Owned by: | Martyn Gigg |
|---|---|---|---|
| Priority: | critical | Milestone: | Iteration 28 |
| Component: | Mantid | Keywords: | |
| Cc: | nick.draper@… | Blocked By: | |
| Blocking: | Tester: | Roman Tolchenov |
Description
With the latest release Winfried from GEM noticed that his analysis takes ~7 mins where it used to take 1 minute. The problem is that it is hitting ManagedWorkspaces when it should not be.
Attachments
Change History
comment:1 Changed 10 years ago by Martyn Gigg
The attached script demonstrates the problem on a Windows 32-biit machine. The input raw files are in
ISIS\ndxgem\Instrument\data\cycle_10_2 and the processed nexus file is attached to the ticket as well.
NOTE: The same NeXus file is loaded 6 times just for easy of carrying files around. The true reduction uses 6 separate files but each is the same size, which is what we care about here.
comment:2 Changed 10 years ago by Martyn Gigg
- Cc nick.draper@… added
- Priority changed from major to critical
- Status changed from new to accepted
comment:3 Changed 10 years ago by Martyn Gigg
(In [9740]) Refs #2568. Fixed the reserved memory calculation on Windows which was return 0. This meant that a workspace went managed even though the reserved memory could have saved it. Also re-enabled the low-fragmentation heap and mmap schemes which weren't being enabled in the right place. I've actually moved them to when the FrameworkManager starts up so that it happens as early as possible.
comment:6 Changed 10 years ago by Martyn Gigg
- Status changed from accepted to verify
- Resolution set to fixed
