perf: remove useless reserve

[yannham]

After looking at flamegraphs for other reasons, some reserves took a surprisingly visible portion of the graph (on the order of the %). Those reserve don't make sense (they unconditionally reserve more space for cases where it's not clear what will happen, which is a waste if we don't need it, and in the event where we actually extend the vector, the default behavior should only resize once in the vast majority of cases).

The most surprising effect is on memory though. This seems to have a large impact, with overall 22% less peak memory consumption on the OPL benchmark, and 27% less leaked memory in total.

[github-actions]

Bencher Report

Branch	perf/remove-useless-reserve
Testbed	ubuntu-latest

Click to view all benchmark results

Benchmark	Latency	microseconds (µs)
diagnostics-benches/inputs/goto-perf.ncl	📈 view plot 🚷 view threshold	10,241.00 µs
diagnostics-benches/inputs/large-record-tree.ncl	📈 view plot 🚷 view threshold	191,300.00 µs
diagnostics-benches/inputs/redis-replication-controller.ncl	📈 view plot 🚷 view threshold	307.94 µs
diagnostics-benches/inputs/small-record-tree.ncl	📈 view plot 🚷 view threshold	429.74 µs
fibonacci 10	📈 view plot 🚷 view threshold	220.95 µs
foldl arrays 50	📈 view plot 🚷 view threshold	393.44 µs
foldl arrays 500	📈 view plot 🚷 view threshold	2,825.30 µs
foldr strings 50	📈 view plot 🚷 view threshold	2,882.30 µs
foldr strings 500	📈 view plot 🚷 view threshold	27,284.00 µs
generate normal 250	📈 view plot 🚷 view threshold	22,944.00 µs
generate normal 50	📈 view plot 🚷 view threshold	1,176.20 µs
generate normal unchecked 1000	📈 view plot 🚷 view threshold	1,885.40 µs
generate normal unchecked 200	📈 view plot 🚷 view threshold	381.04 µs
init-diagnostics-benches/inputs/goto-perf.ncl	📈 view plot 🚷 view threshold	53,559.00 µs
init-diagnostics-benches/inputs/large-record-tree.ncl	📈 view plot 🚷 view threshold	207,040.00 µs
init-diagnostics-benches/inputs/redis-replication-controller.ncl	📈 view plot 🚷 view threshold	48,850.00 µs
init-diagnostics-benches/inputs/small-record-tree.ncl	📈 view plot 🚷 view threshold	48,652.00 µs
pidigits 100	📈 view plot 🚷 view threshold	1,882.70 µs
pipe normal 20	📈 view plot 🚷 view threshold	607.47 µs
pipe normal 200	📈 view plot 🚷 view threshold	4,491.70 µs
product 30	📈 view plot 🚷 view threshold	264.39 µs
requests-benches/inputs/goto-perf.ncl-000	📈 view plot 🚷 view threshold	3,035.50 µs
requests-benches/inputs/large-record-tree.ncl-000	📈 view plot 🚷 view threshold	574,310.00 µs
requests-benches/inputs/large-record-tree.ncl-001	📈 view plot 🚷 view threshold	86.02 µs
scalar 10	📈 view plot 🚷 view threshold	492.90 µs
sum 30	📈 view plot 🚷 view threshold	264.89 µs

🐰 View full continuous benchmarking report in Bencher

[jneem]

Nice find!