[SERVER-21008] 20% Single thread performance regression on Insert.EmptyCapped and Insert.CappedSeqIntID on wiredTiger Created: 08/Oct/15 Updated: 01/Apr/16 Resolved: 30/Oct/15 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Performance, WiredTiger |
| Affects Version/s: | None |
| Fix Version/s: | 3.2.0-rc2 |
| Type: | Bug | Priority: | Major - P3 |
| Reporter: | David Daly | Assignee: | Martin Bligh |
| Resolution: | Done | Votes: | 0 |
| Labels: | mpreg | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||
| Backwards Compatibility: | Fully Compatible | ||||
| Operating System: | ALL | ||||
| Participants: | |||||
| Linked BF Score: | 0 | ||||
| Description |
|
20% regression in single thread performance compared to baseline. This is using a recent compare to 3.1.7. To recreate the empty capped in the shell:
Original failure message:
|
| Comments |
| Comment by Githook User [ 30/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Author: {u'username': u'martinbligh', u'name': u'Martin Bligh', u'email': u'mbligh@mongodb.com'}Message: | ||||||||||||||||||||||||||||||||||||||||
| Comment by Githook User [ 29/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Author: {u'username': u'martinbligh', u'name': u'Martin Bligh', u'email': u'mbligh@mongodb.com'}Message: | ||||||||||||||||||||||||||||||||||||||||
| Comment by Githook User [ 29/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Author: {u'username': u'martinbligh', u'name': u'Martin Bligh', u'email': u'mbligh@mongodb.com'}Message: Revert " This reverts commit 9d7413280fb706e037082ea84442bb83579df58c. | ||||||||||||||||||||||||||||||||||||||||
| Comment by Githook User [ 29/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Author: {u'username': u'martinbligh', u'name': u'Martin Bligh', u'email': u'mbligh@mongodb.com'}Message: | ||||||||||||||||||||||||||||||||||||||||
| Comment by Ramon Fernandez Marina [ 26/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Githook message for commit to this ticket:
| ||||||||||||||||||||||||||||||||||||||||
| Comment by Martin Bligh [ 19/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Forward porting that change fixed the caller list but doesn't seem to have improved perf at all:
| ||||||||||||||||||||||||||||||||||||||||
| Comment by Martin Bligh [ 19/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Most obvious difference seems to be a lot more time spent in __wt_btcur_next. 3.0.6
3.2.0-rc0
| ||||||||||||||||||||||||||||||||||||||||
| Comment by David Daly [ 16/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Link to result page comparing to 3.1.7. Should be in the graph view. Look at the 1 and 2 thread bars. | ||||||||||||||||||||||||||||||||||||||||
| Comment by David Daly [ 16/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
The issue doesn't show up with 8 threads. It only fails at 1 and 2 threads, but is 20% for those cases. | ||||||||||||||||||||||||||||||||||||||||
| Comment by Martin Bligh [ 15/Oct/15 ] | ||||||||||||||||||||||||||||||||||||||||
|
Tried rerunning Insert.EmptyCapped locally w/8 threads & 30 s elapsed. 3.0.6 master So if anything it seems faster. |