-
Notifications
You must be signed in to change notification settings - Fork 17.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
bytes: speed up Replace #72967
base: master
Are you sure you want to change the base?
bytes: speed up Replace #72967
Conversation
This PR (HEAD: ca56e09) has been imported to Gerrit for code review. Please visit Gerrit at https://go-review.googlesource.com/c/go/+/659515. Important tips:
|
Message from Gopher Robot: Patch Set 1: (1 comment) Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
Message from jub0bs: Patch Set 1: (1 comment) Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
Message from Jorropo: Patch Set 1: Commit-Queue+1 Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
Message from Go LUCI: Patch Set 1: Dry run: CV is trying the patch. Bot data: {"action":"start","triggered_at":"2025-03-20T15:04:30Z","revision":"fa97c61fc9da668b2702c88d2087cb9c56538350"} Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
Message from Jorropo: Patch Set 1: -Commit-Queue Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
Message from Go LUCI: Patch Set 1: This CL has passed the run Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
Message from Go LUCI: Patch Set 1: LUCI-TryBot-Result+1 Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
This CL is to package bytes what CL 657935 was to package strings. The length of parameter old does not change. Move the corresponding length check outside the loop. Use range-over-int loops where possible. Some benchmark results (no changes to allocations): goos: darwin goarch: amd64 pkg: bytes cpu: Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz │ old │ new │ │ sec/op │ sec/op vs base │ Replace/"hello"_"l"_"L"_0-8 27.83n ± 2% 27.22n ± 1% -2.17% (p=0.000 n=20) Replace/"hello"_"l"_"L"_-1-8 60.64n ± 0% 57.97n ± 1% -4.40% (p=0.000 n=20) Replace/"hello"_"x"_"X"_-1-8 33.99n ± 0% 33.88n ± 0% ~ (p=0.140 n=20) Replace/""_"x"_"X"_-1-8 10.40n ± 1% 10.57n ± 0% +1.64% (p=0.000 n=20) Replace/"radar"_"r"_"<r>"_-1-8 62.63n ± 0% 61.39n ± 0% -1.98% (p=0.000 n=20) Replace/""_""_"<>"_-1-8 29.76n ± 1% 24.18n ± 1% -18.75% (p=0.000 n=20) Replace/"banana"_"a"_"<>"_-1-8 77.00n ± 0% 77.10n ± 1% ~ (p=0.525 n=20) Replace/"banana"_"a"_"<>"_1-8 44.24n ± 0% 43.57n ± 1% -1.54% (p=0.000 n=20) Replace/"banana"_"a"_"<>"_1000-8 78.23n ± 0% 77.16n ± 1% -1.36% (p=0.000 n=20) Replace/"banana"_"an"_"<>"_-1-8 72.78n ± 1% 69.97n ± 1% -3.85% (p=0.000 n=20) Replace/"banana"_"ana"_"<>"_-1-8 54.41n ± 0% 54.04n ± 1% -0.67% (p=0.033 n=20) Replace/"banana"_""_"<>"_-1-8 116.8n ± 1% 103.5n ± 1% -11.42% (p=0.000 n=20) Replace/"banana"_""_"<>"_10-8 117.2n ± 1% 103.6n ± 0% -11.60% (p=0.000 n=20) Replace/"banana"_""_"<>"_6-8 105.30n ± 0% 92.50n ± 0% -12.16% (p=0.000 n=20) Replace/"banana"_""_"<>"_5-8 91.81n ± 0% 79.87n ± 1% -13.01% (p=0.000 n=20) Replace/"banana"_""_"<>"_1-8 35.87n ± 1% 30.33n ± 1% -15.43% (p=0.000 n=20) Replace/"banana"_"a"_"a"_-1-8 70.84n ± 0% 68.83n ± 0% -2.84% (p=0.000 n=20) Replace/"banana"_"a"_"a"_1-8 44.27n ± 1% 43.47n ± 1% -1.80% (p=0.000 n=20) Replace/"☺☻☹"_""_"<>"_-1-8 104.25n ± 1% 93.33n ± 0% -10.48% (p=0.000 n=20) geomean 56.31n 52.88n -6.09%
ca56e09
to
d1caf1f
Compare
This PR (HEAD: d1caf1f) has been imported to Gerrit for code review. Please visit Gerrit at https://go-review.googlesource.com/c/go/+/659515. Important tips:
|
Message from jub0bs: Patch Set 1: (1 comment) Please don’t reply on this GitHub thread. Visit golang.org/cl/659515. |
This CL is to package bytes what CL 657935 was to package strings.
Add fuzz test and benchmarks for Replace.
The length of parameter old does not change. Move the corresponding length
check outside the loop. Use range-over-int loops where possible.
Some benchmark results (no changes to allocations):
goos: darwin
goarch: amd64
pkg: bytes
cpu: Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz
│ old │ new │
│ sec/op │ sec/op vs base │
Replace/"hello""l""L"0-8 27.83n ± 2% 27.22n ± 1% -2.17% (p=0.000 n=20)
Replace/"hello""l""L"-1-8 60.64n ± 0% 57.97n ± 1% -4.40% (p=0.000 n=20)
Replace/"hello""x""X"-1-8 33.99n ± 0% 33.88n ± 0% ~ (p=0.140 n=20)
Replace/"""x""X"-1-8 10.40n ± 1% 10.57n ± 0% +1.64% (p=0.000 n=20)
Replace/"radar""r"""-1-8 62.63n ± 0% 61.39n ± 0% -1.98% (p=0.000 n=20)
Replace/"""""<>"-1-8 29.76n ± 1% 24.18n ± 1% -18.75% (p=0.000 n=20)
Replace/"banana""a""<>"-1-8 77.00n ± 0% 77.10n ± 1% ~ (p=0.525 n=20)
Replace/"banana""a""<>"1-8 44.24n ± 0% 43.57n ± 1% -1.54% (p=0.000 n=20)
Replace/"banana""a""<>"1000-8 78.23n ± 0% 77.16n ± 1% -1.36% (p=0.000 n=20)
Replace/"banana""an""<>"-1-8 72.78n ± 1% 69.97n ± 1% -3.85% (p=0.000 n=20)
Replace/"banana""ana""<>"-1-8 54.41n ± 0% 54.04n ± 1% -0.67% (p=0.033 n=20)
Replace/"banana""""<>"-1-8 116.8n ± 1% 103.5n ± 1% -11.42% (p=0.000 n=20)
Replace/"banana""""<>"10-8 117.2n ± 1% 103.6n ± 0% -11.60% (p=0.000 n=20)
Replace/"banana""""<>"6-8 105.30n ± 0% 92.50n ± 0% -12.16% (p=0.000 n=20)
Replace/"banana""""<>"5-8 91.81n ± 0% 79.87n ± 1% -13.01% (p=0.000 n=20)
Replace/"banana""""<>"1-8 35.87n ± 1% 30.33n ± 1% -15.43% (p=0.000 n=20)
Replace/"banana""a""a"-1-8 70.84n ± 0% 68.83n ± 0% -2.84% (p=0.000 n=20)
Replace/"banana""a""a"1-8 44.27n ± 1% 43.47n ± 1% -1.80% (p=0.000 n=20)
Replace/"☺☻☹""""<>"_-1-8 104.25n ± 1% 93.33n ± 0% -10.48% (p=0.000 n=20)
geomean 56.31n 52.88n -6.09%