Benchmarks of pull and push iterators in Go.

README.md

A simple benchmark comparing push and pull iterators to each other and also comparing them to themselves with and without PGO.

Results

goos: linux
goarch: amd64
pkg: test
cpu: AMD Ryzen 9 3900X 12-Core Processor
        │ no-pgo.txt  │              pgo.txt               │
        │   sec/op    │   sec/op     vs base               │
Push-24   116.5n ± 6%   107.4n ± 2%  -7.77% (p=0.000 n=10)
Pull-24   2.047µ ± 2%   1.994µ ± 7%  -2.57% (p=0.009 n=10)
geomean   488.2n        462.8n       -5.20%

        │ no-pgo.txt │               pgo.txt               │
        │    B/op    │    B/op     vs base                 │
Push-24   24.00 ± 0%   24.00 ± 0%       ~ (p=1.000 n=10) ¹
Pull-24   352.0 ± 0%   352.0 ± 0%       ~ (p=1.000 n=10) ¹
geomean   91.91        91.91       +0.00%
¹ all samples are equal

        │ no-pgo.txt │               pgo.txt               │
        │ allocs/op  │ allocs/op   vs base                 │
Push-24   2.000 ± 0%   2.000 ± 0%       ~ (p=1.000 n=10) ¹
Pull-24   12.00 ± 0%   12.00 ± 0%       ~ (p=1.000 n=10) ¹
geomean   4.899        4.899       +0.00%
¹ all samples are equal

go.mod

module test

go 1.23.0

iter_test.go

package iter_test

import (
	"iter"
	"slices"
	"testing"
)

var values = slices.Values([]int{1, 2, 3, 4, 5, 6, 7, 8, 9, 10})

func BenchmarkPush(b *testing.B) {
	for range b.N {
		for v := range values {
			_ = v
		}
	}
}

func BenchmarkPull(b *testing.B) {
	for range b.N {
		next, stop := iter.Pull(values)
		for {
			_, ok := next()
			if !ok {
				break
			}
		}
		stop()
	}
}