Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Small CPU optimization in InlineSkipList::Insert() (facebook#12975)
Summary: reuse decode key in more places to avoid decoding length prefixed key x->Key(). Pull Request resolved: facebook#12975 Test Plan: ran benchmarks simultaneously for "before" and "after" * fillseq: ``` (for I in $(seq 1 50); do ./db_bench --benchmarks=fillseq --disable_auto_compactions=1 --min_write_buffer_number_to_merge=100 --max_write_buffer_number=1000 --write_buffer_size=268435456 --num=5000000 --seed=1723056275 --disable_wal=1 2>&1 | grep "fillseq" done;) | awk '{ t += $5; c++; print } END { printf ("%9.3f\n", 1.0 * t / c) }'; before: 1483191 after: 1490555 (+0.5%) ``` * fillrandom: ``` (for I in $(seq 1 2); do ./db_bench_imain --benchmarks=fillrandom --disable_auto_compactions=1 --min_write_buffer_number_to_merge=100 --max_write_buffer_number=1000 --write_buffer_size=268435456 --num=2500000 --seed=1723056275 --disable_wal=1 2>&1 | grep "fillrandom" before: 255463 after: 256128 (+0.26%) ``` Reviewed By: anand1976 Differential Revision: D61835340 Pulled By: cbi42 fbshipit-source-id: 70345510720e348bacd51269acb5d2dd5a62bf0a
- Loading branch information