Optimized set type random performance and cleanup list code #1406
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## unstable #1406 +/- ##
============================================
- Coverage 70.82% 70.82% -0.01%
============================================
Files 118 118
Lines 63561 63555 -6
============================================
- Hits 45020 45014 -6
Misses 18541 18541
|
src/unit/test_listpack.c
Outdated
| for (size_t i = 0; i < count; i++) { | ||
| unsigned int j = 0; | ||
| // unsigned int len = 0; /* initialize to silence warning */ | ||
| // long long llele = 0; /* initialize to silence warning */ |
There was a problem hiding this comment.
Unused stuff left from debugging?
Signed-off-by: artikell <[email protected]>
artikell
force-pushed
the
opti_set_random
branch
from
2963107 to
76bbcf6
Compare
|
I'm sorry for this MR. After repeated testing, I found that this was a negative optimization, and the previous conclusion was that there was a problem with the time count. lpNextRandom, List size: 1, total 10208 usec Here is a new result, lpSeek performs better, which also implies that lpNextRandom may be an optimizable function. I decided to close this MR. Follow up and think about how to optimize. |
There are currently two random methods, and the random+lseek method has poor performance.
Test result(test_listpackBenchRandom)：
lpNextRandom, List size: 10000, average 58 usec
lpSeek, List size: 10000, average 86 usec