strings: performance regression in strings/CountHard* benchmarks

[archanaravindar]

Go version

Go1.27rc2, go version go1.28-devel_5cbe9d2fe2

Output of go env in your module/workspace:

GO111MODULE=''
GOAMD64='v1'
GOARCH='amd64'
GOAUTH='netrc'
GOBIN='/home/archanaravindar/Go/Go_master/go/bin'
GOCACHE='/home/archanaravindar/.cache/go-build'
GOCACHEPROG=''
GODEBUG=''
GOENV='/home/archanaravindar/.config/go/env'
GOEXE=''
GOEXPERIMENT='nodwarf5'
GOFIPS140='off'
GOFLAGS=''
GOGCCFLAGS='-fPIC -m64 -pthread -Wl,--no-gc-sections -fmessage-length=0 -ffile-prefix-map=/tmp/go-build2580094643=/tmp/go-build -gno-record-gcc-switches'
GOHOSTARCH='amd64'
GOHOSTOS='linux'
GOINSECURE=''
GOMOD='/home/archanaravindar/upstream/counthard/go/src/go.mod'
GOMODCACHE='/home/archanaravindar/go/pkg/mod'
GONOPROXY=''
GONOSUMDB=''
GOOS='linux'
GOPATH='/home/archanaravindar/go'
GOPRIVATE=''
GOPROXY='https://proxy.golang.org,direct'
GOROOT='/usr/lib/golang'
GOSUMDB='sum.golang.org'
GOTELEMETRY='local'
GOTELEMETRYDIR='/home/archanaravindar/.config/go/telemetry'
GOTMPDIR=''
GOTOOLCHAIN='local'
GOTOOLDIR='/usr/lib/golang/pkg/tool/linux_amd64'
GOVCS=''
GOVERSION='go1.25.12 X:nodwarf5'

What did you do?

Ran countHard tests in string package

What did you see happen?

Saw a performance regression on amd64 and ppc64le
(1.27rc2 vs 1.26.5 and master vs 1.26.5)

goos: linux
goarch: amd64
pkg: strings
cpu: 11th Gen Intel(R) Core(TM) i7-1185G7 @ 3.00GHz
             │   base.out   │ ../../../go1.27rc2/src/strings/new.out  │
             │    sec/op    │    sec/op      vs base                  │
CountHard1-8   90.51µ ±  4%   1172.52µ ± 5%  +1195.47% (p=0.000 n=10)
CountHard2-8   132.3µ ± 16%    1243.5µ ± 5%   +840.00% (p=0.000 n=10)
CountHard3-8   427.5µ ±  6%    1222.9µ ± 7%   +186.04% (p=0.000 n=10)
geomean        172.3µ           1.213m        +603.60%

This affects not just CountHard but multiple benchmarks exercising this path

What did you expect to see?

Expected to see similar performance as Go1.26.5

[gopherbot]

Change https://go.dev/cl/807360 mentions this issue: internal/stringslite: fix performance regression in index

[gabyhelp]

Related Issues

• strings: 10-30% speed regression in Contains from 1.13 to tip #35686 (closed)
• strings: IndexAny performance regression vs 1.9 #22750 (closed)

Related Code Changes

• bytes: make Count faster
• internal/bytealg: optimize Index for riscv64
• internal/bytealg: optimize Count/CountString in arm64
• internal/bytealg: optimize Count/CountString in amd64
• internal/bytealg: optimize Count with PCALIGN in riscv64
• strings: optimize Count for amd64
• runtime: implement cmpstring and bytes.Compare in assembly for ppc64
• bytes,strings: use IndexByte more often in Index on AMD64

_{(Emoji vote if this was helpful or unhelpful; more detailed feedback welcome in this discussion.)}

[randall77]

Reopening for 1.27 backport.

[gopherbot]

Change https://go.dev/cl/809800 mentions this issue: [release-branch.go1.27] internal/stringslite: fix performance regression in index

[gopherbot]

Closed by merging CL 809800 (commit 2b10e5a) to release-branch.go1.27.