1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
|
// Copyright 2015, Joe Tsai. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE.md file.
// Benchmark tool to compare performance between multiple compression
// implementations. Individual implementations are referred to as codecs.
//
// Example usage:
//
// $ go build
// $ ./bench \
// -formats fl \
// -tests encRate,decRate \
// -codecs std,ds,cgo \
// -globs twain.txt \
// -levels 1,6,9 \
// -sizes 1e4,1e5,1e6
//
//
// BENCHMARK: fl:encRate
// benchmark std MB/s delta cgo MB/s delta
// twain.txt:1:1e4 9.88 1.00x 48.89 4.95x
// twain.txt:1:1e5 26.70 1.00x 64.99 2.43x
// twain.txt:1:1e6 31.95 1.00x 65.56 2.05x
// twain.txt:6:1e4 7.31 1.00x 30.67 4.19x
// twain.txt:6:1e5 8.33 1.00x 17.22 2.07x
// twain.txt:6:1e6 8.05 1.00x 15.99 1.99x
// twain.txt:9:1e4 8.15 1.00x 30.04 3.69x
// twain.txt:9:1e5 6.59 1.00x 12.82 1.95x
// twain.txt:9:1e6 6.32 1.00x 11.40 1.80x
//
// BENCHMARK: fl:decRate
// benchmark std MB/s delta ds MB/s delta cgo MB/s delta
// twain.txt:1:1e4 49.61 1.00x 74.15 1.49x 163.81 3.30x
// twain.txt:1:1e5 60.25 1.00x 91.25 1.51x 177.38 2.94x
// twain.txt:1:1e6 61.75 1.00x 95.82 1.55x 181.11 2.93x
// twain.txt:6:1e4 52.16 1.00x 77.25 1.48x 174.30 3.34x
// twain.txt:6:1e5 72.23 1.00x 108.01 1.50x 195.31 2.70x
// twain.txt:6:1e6 76.59 1.00x 116.80 1.53x 203.88 2.66x
// twain.txt:9:1e4 52.97 1.00x 77.58 1.46x 172.88 3.26x
// twain.txt:9:1e5 72.35 1.00x 108.37 1.50x 197.15 2.72x
// twain.txt:9:1e6 76.82 1.00x 118.02 1.54x 204.87 2.67x
//
//
// RUNTIME: 2m42.434570856s
package main
import (
"flag"
"fmt"
"math"
"os"
"path/filepath"
"strings"
"time"
)
func main() {
flag.CommandLine.Usage = func() {
fmt.Fprintf(os.Stderr, "Usage of %s:\n", os.Args[0])
flag.CommandLine.VisitAll(func(f *flag.Flag) {
if strings.HasPrefix(f.Name, "test.") {
return
}
typ, usage := flag.UnquoteUsage(f)
s := fmt.Sprintf(" -%s %s\n %s (default %q)\n",
f.Name, typ, usage, f.DefValue)
fmt.Fprint(os.Stderr, s)
})
}
setDefaults()
flag.Var(&formats, "formats", "List of formats to benchmark")
flag.Var(&tests, "tests", "List of different benchmark tests")
flag.Var(&codecs, "codecs", "List of codecs to benchmark")
flag.Var(&paths, "paths", "List of paths to search for test files")
flag.Var(&globs, "globs", "List of globs to match for test files")
flag.Var(&levels, "levels", "List of compression levels to benchmark")
flag.Var(&sizes, "sizes", "List of input sizes to benchmark")
flag.Parse()
files := getFiles(paths, globs)
ts := time.Now()
runBenchmarks(files, codecs, formats, tests, levels, sizes)
te := time.Now()
fmt.Printf("RUNTIME: %v\n", te.Sub(ts))
}
type file struct{ Abs, Rel string }
// getFiles returns a list of files found by applying the glob matching on
// all the specified paths. This function ignores any errors.
func getFiles(paths []string, globs []string) []file {
var fs []file
for _, p := range paths {
for _, g := range globs {
ms, _ := filepath.Glob(filepath.Join(p, g))
for _, m := range ms {
r, err1 := filepath.Rel(p, m)
fi, err2 := os.Stat(m)
if err1 == nil && err2 == nil && !fi.IsDir() {
fs = append(fs, file{Abs: m, Rel: r})
}
}
}
}
return fs
}
func runBenchmarks(files []file, codecs []string, formats []Format, tests []Test, levels, sizes []int) {
for _, f := range formats {
// Get lists of encoders and decoders that exist.
var encs, decs []string
for _, c := range codecs {
if _, ok := encoders[f][c]; ok {
encs = append(encs, c)
}
}
for _, c := range codecs {
if _, ok := decoders[f][c]; ok {
decs = append(decs, c)
}
}
for _, t := range tests {
var results [][]Result
var names, codecs []string
var title, suffix string
// Check that we can actually do this
fmt.Printf("BENCHMARK: %s:%s\n", enumToFmt[f], enumToTest[t])
if len(encs) == 0 {
fmt.Println("\tSKIP: There are no encoders available.")
fmt.Println("")
continue
}
if len(decs) == 0 && t == TestDecodeRate {
fmt.Println("\tSKIP: There are no decoders available.")
fmt.Println("")
continue
}
// Progress ticker.
var cnt int
tick := func() {
total := len(codecs) * len(files) * len(levels) * len(sizes)
pct := 100.0 * float64(cnt) / float64(total)
fmt.Printf("\t[%6.2f%%] %d of %d\r", pct, cnt, total)
cnt++
}
// Perform the This may take some time.
switch t {
case TestEncodeRate:
codecs, title, suffix = encs, "MB/s", ""
results, names = BenchmarkEncoderSuite(f, encs, files, levels, sizes, tick)
case TestDecodeRate:
ref := getReferenceEncoder(f)
codecs, title, suffix = decs, "MB/s", ""
results, names = BenchmarkDecoderSuite(f, decs, files, levels, sizes, ref, tick)
case TestCompressRatio:
codecs, title, suffix = encs, "ratio", "x"
results, names = BenchmarkRatioSuite(f, encs, files, levels, sizes, tick)
default:
panic("unknown test")
}
// Print all of the results.
printResults(results, names, codecs, title, suffix)
fmt.Println()
}
fmt.Println()
}
}
// The decompression speed benchmark works by decompressing some pre-compressed
// data. In order for the benchmarks to be consistent, the same encoder should
// be used to generate the pre-compressed data for all the trials.
//
// encRefs defines the priority order for which encoders to choose first as the
// reference compressor. If no compressor is found for any of the listed codecs,
// then a random encoder will be chosen.
var encRefs = []string{"std", "cgo", "ds"}
func getReferenceEncoder(f Format) Encoder {
for _, c := range encRefs {
if enc, ok := encoders[f][c]; ok {
return enc // Choose by priority
}
}
for _, enc := range encoders[f] {
return enc // Choose any random encoder
}
return nil // There are no encoders
}
func printResults(results [][]Result, names, codecs []string, title, suffix string) {
// Allocate result table.
cells := make([][]string, 1+len(names))
for i := range cells {
cells[i] = make([]string, 1+2*len(codecs))
}
// Label the first row.
cells[0][0] = "benchmark"
for i, c := range codecs {
cells[0][1+2*i] = c + " " + title
cells[0][2+2*i] = "delta"
}
// Insert all rows.
for j, row := range results {
cells[1+j][0] = names[j]
for i, r := range row {
if r.R != 0 && !math.IsNaN(r.R) && !math.IsInf(r.R, 0) {
cells[1+j][1+2*i] = fmt.Sprintf("%.2f", r.R) + suffix
}
if r.D != 0 && !math.IsNaN(r.D) && !math.IsInf(r.D, 0) {
cells[1+j][2+2*i] = fmt.Sprintf("%.2f", r.D) + "x"
}
}
}
// Compute the maximum lengths.
maxLens := make([]int, 1+2*len(codecs))
for _, row := range cells {
for i, s := range row {
if maxLens[i] < len(s) {
maxLens[i] = len(s)
}
}
}
// Print padded versions of all cells.
for _, row := range cells {
fmt.Print("\t")
for i, s := range row {
switch {
case i == 0: // Column 0
row[i] = s + strings.Repeat(" ", maxLens[i]-len(s))
case i%2 == 1: // Column 1, 3, 5, 7, ...
row[i] = strings.Repeat(" ", 6+maxLens[i]-len(s)) + s
case i%2 == 0: // Column 2, 4, 6, 8, ...
row[i] = strings.Repeat(" ", 2+maxLens[i]-len(s)) + s
}
fmt.Print(row[i])
}
fmt.Println()
}
}
|