66 lines
1.7 KiB
Go
66 lines
1.7 KiB
Go
// Copyright (c) 2026 Petr Balvín <opensource@petrbalvin.org> (https://petrbalvin.org)
|
|
// SPDX-License-Identifier: MIT
|
|
|
|
package io
|
|
|
|
// The A/B benchmark for the CSV read path: the legacy encoding/csv
|
|
// reader and the hand-rolled tokenizer side by side, in one process.
|
|
// The two sub-benchmarks alternate within every -count round, so both
|
|
// see the same machine, and the decision reads the medians with 1 to 2
|
|
// percent treated as noise.
|
|
|
|
import (
|
|
"bytes"
|
|
"io"
|
|
"testing"
|
|
|
|
"sourcedock.dev/petrbalvin/tensor/internal/core"
|
|
)
|
|
|
|
// The A/B table: a hundred thousand rows of sixteen float64 values, in
|
|
// the text shape SaveCSV emits, sized so the per-value work dominates
|
|
// everything else.
|
|
const (
|
|
benchCSVABRows = 100000
|
|
benchCSVABCols = 16
|
|
)
|
|
|
|
// benchCSVABData renders the A/B table as CSV bytes, once, before the
|
|
// measured loops.
|
|
func benchCSVABData(b *testing.B) []byte {
|
|
b.Helper()
|
|
var buf bytes.Buffer
|
|
if err := SaveCSVWriter(&buf, benchFloats(benchCSVABRows, benchCSVABCols)); err != nil {
|
|
b.Fatal(err)
|
|
}
|
|
return buf.Bytes()
|
|
}
|
|
|
|
// BenchmarkLoadCSVReader races the two readers over the same table.
|
|
func BenchmarkLoadCSVReader(b *testing.B) {
|
|
data := benchCSVABData(b)
|
|
impls := []struct {
|
|
name string
|
|
load func(io.Reader, bool) (*core.Array, error)
|
|
}{
|
|
{"old", loadCSVReaderLegacy},
|
|
{"new", LoadCSVReader},
|
|
}
|
|
for _, impl := range impls {
|
|
b.Run(impl.name, func(b *testing.B) {
|
|
r := bytes.NewReader(data)
|
|
b.ReportAllocs()
|
|
for b.Loop() {
|
|
r.Reset(data)
|
|
a, err := impl.load(r, false)
|
|
if err != nil {
|
|
b.Fatal(err)
|
|
}
|
|
if a.Shape()[0] != benchCSVABRows || a.Shape()[1] != benchCSVABCols {
|
|
b.Fatalf("array came back with shape %v", a.Shape())
|
|
}
|
|
}
|
|
})
|
|
}
|
|
}
|