You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Mar 22, 2023. It is now read-only.
In the default benchmark execution order, deleterandom is executed after deleteseq. This gives very low numbers for deleterandom, which can be explained in tree-based engines by less traversal due to previous deletions. Should deleterandom/deleteseq be executed just after fillseq (anyhow db is reset before fillrandom)?
The whole idea assumes default execution order, with no --benchmarks parameter supplied.
The text was updated successfully, but these errors were encountered:
Well, there is probably no one, best ordering for those benchmarks. The performance of delete (both seq and random) will probably be different depending on whether you called fillrandom or fillseq before.
I believe that you should always specify the benchmarks explicitly (run them in a specific order) - to best simulate your workload.
In the default benchmark execution order, deleterandom is executed after deleteseq. This gives very low numbers for deleterandom, which can be explained in tree-based engines by less traversal due to previous deletions. Should deleterandom/deleteseq be executed just after fillseq (anyhow db is reset before fillrandom)?
The whole idea assumes default execution order, with no --benchmarks parameter supplied.
The text was updated successfully, but these errors were encountered: