| Age | Commit message (Collapse) | Author | Files | |
|---|---|---|---|---|
| 2021-11-14 | Don't hardcode target gdbm versionsHEADmaster | Sergey Poznyakoff | 7 | |
| 2020-10-29 | runtest: Set initial cache capacity to 1 | Sergey Poznyakoff | 1 | |
| 2020-10-29 | Minor changes | Sergey Poznyakoff | 3 | |
| * Makefile.am (benchmark): New variable FETCHKEYSOPT. * gdbm/newcache: Upgrade. * src/dropcache.c: Call sync before dropping caches. * src/fetchkeys.c: Use CPU user and system times, instead of measuring real execution time. Don't count gdbm_close in the totals. * src/gnuplot.m4: Use bezier instead of csplines. | ||||
| 2020-03-17 | Improve fetchkeys | Sergey Poznyakoff | 2 | |
| The new option -b instructs it to fetch only the bucket for each given key without fetching the actual value, thus eliminating the imact of non-cached reads. * src/benchmark.mk.in: Pass FETCHKEYSOPT to fetchkeys. * src/fetchkeys.c: New option -b: read the bucket, don't fetch value | ||||
| 2020-03-15 | Use actual min and max times for errorbar. Limit the maxcache value by the ↵ | Sergey Poznyakoff | 3 | |
| number of buckets in the db plus 100. | ||||
| 2020-03-15 | Improve data collection and the gnuplot script. | Sergey Poznyakoff | 3 | |
| 2020-03-15 | Initial commit | Sergey Poznyakoff | 9 | |
 |
index : gdbm/cache-benchmarks.git | |
| Comparative benchmarks of GDBM cache performance | Sergey Poznyakoff |
| aboutsummaryrefslogtreecommitdiff |
Return to:
Send suggestions and report system problems to the System administrator.