Viewing all changes in revision 4086.
- Committer: Tarmac
- Date: 2017-03-15 08:43:38 UTC
- mfrom: (4084.2.3 medi)
- Revision ID: tarmac-20170315084338-24ypetor0hll2qgt
Optimize a crazily inefficient test case (LP: #1671033)
Not a threading problem this time, but a general algorithmic complexity
problem. The test itself was running 50x longer than it needs to, and
the FakeAlarmFactory class it uses was copying a vector 3 million times during the one test case (now zero times).
Execution times under valgrind:
Before: 30 seconds; After: 0.7 seconds
Execution times natively:
Before: 0.8 seconds; After: 0.01 seconds. Fixes: https://bugs.launchpad.net/bugs/1671033.
Approved by Alan Griffiths, mir-ci-bot.
Not a threading problem this time, but a general algorithmic complexity
problem. The test itself was running 50x longer than it needs to, and
the FakeAlarmFactory class it uses was copying a vector 3 million times during the one test case (now zero times).
Execution times under valgrind:
Before: 30 seconds; After: 0.7 seconds
Execution times natively:
Before: 0.8 seconds; After: 0.01 seconds. Fixes: https://bugs.launchpad.net/bugs/1671033.
Approved by Alan Griffiths, mir-ci-bot.
expand all
collapse all
added
removed
