< section intentionally left blank >
I initially wrote this a very bad way (yeah, I knew it while it was happening but just kept moving forward). It was fairly easy to optimize, so I kept both versions and used benchstat to compare the outputs of go.exe test -bench=. -run='^$' -tags test -count=10 -benchmem ./go:
goos: windows
goarch: amd64
pkg: github.com/asphaltbuffet/advent-of-code/exercises/2023/06-waitForIt/go
cpu: Intel(R) Core(TM) i7-9700K CPU @ 3.60GHz
│ old.bench │ new.bench │
│ sec/op │ sec/op vs base │
PartOne/input_0-8 1396.0n ± 2% 890.2n ± 2% -36.23% (p=0.000 n=10)
PartOne/input_1-8 3.648µ ± 4% 1.127µ ± 2% -69.12% (p=0.000 n=10)
PartOne/input_2-8 3.611µ ± 2% 1.128µ ± 2% -68.75% (p=0.000 n=10)
PartTwo/input_0-8 230895.5n ± 9% 507.5n ± 2% -99.78% (p=0.000 n=10)
PartTwo/input_1-8 110945126.0n ± 7% 676.9n ± 1% -100.00% (p=0.000 n=10)
PartTwo/input_2-8 75850688.5n ± 1% 674.8n ± 1% -100.00% (p=0.000 n=10)
geomean 181.5µ 800.1n -99.56%
│ old.bench │ new.bench │
│ B/op │ B/op vs base │
PartOne/input_0-8 2000.0 ± 0% 736.0 ± 0% -63.20% (p=0.000 n=10)
PartOne/input_1-8 8368.0 ± 0% 912.0 ± 0% -89.10% (p=0.000 n=10)
PartOne/input_2-8 7920.0 ± 0% 912.0 ± 0% -88.48% (p=0.000 n=10)
PartTwo/input_0-8 1294514.0 ± 0% 176.0 ± 0% -99.99% (p=0.000 n=10)
PartTwo/input_1-8 1094566079.5 ± 0% 184.0 ± 0% -100.00% (p=0.000 n=10)
PartTwo/input_2-8 751993030.0 ± 0% 184.0 ± 0% -100.00% (p=0.000 n=10)
geomean 704.7Ki 392.3 -99.95%
│ old.bench │ new.bench │
│ allocs/op │ allocs/op vs base │
PartOne/input_0-8 22.00 ± 0% 16.00 ± 0% -27.27% (p=0.000 n=10)
PartOne/input_1-8 27.00 ± 0% 19.00 ± 0% -29.63% (p=0.000 n=10)
PartOne/input_2-8 27.00 ± 0% 19.00 ± 0% -29.63% (p=0.000 n=10)
PartTwo/input_0-8 9.000 ± 0% 7.000 ± 0% -22.22% (p=0.000 n=10)
PartTwo/input_1-8 9.000 ± 0% 7.000 ± 0% -22.22% (p=0.000 n=10)
PartTwo/input_2-8 9.000 ± 0% 7.000 ± 0% -22.22% (p=0.000 n=10)
geomean 15.07 11.21 -25.61%
< section intentionally left blank >
