protobuf-go/internal/impl/codec_map.go

249 lines
6.5 KiB
Go
Raw Normal View History

internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
// Copyright 2019 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package impl
import (
"fmt"
"reflect"
"sort"
"google.golang.org/protobuf/internal/encoding/wire"
"google.golang.org/protobuf/proto"
pref "google.golang.org/protobuf/reflect/protoreflect"
)
var protoMessageType = reflect.TypeOf((*proto.Message)(nil)).Elem()
internal/impl: add fast-path unmarshal Benchmarks run with: go test ./benchmarks/ -bench=Wire -benchtime=500ms -benchmem -count=8 Fast-path vs. parent commit: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 1.35µs ± 2% 0.45µs ± 4% -67.01% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 1.07µs ± 1% 0.31µs ± 1% -71.04% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 691µs ± 2% 188µs ± 2% -72.78% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 60.0 ± 0% 25.0 ± 0% -58.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 42.0 ± 0% 7.0 ± 0% -83.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 28.6k ± 0% 8.5k ± 0% -70.34% (p=0.000 n=8+8) Fast-path vs. -v1: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 702ns ± 1% 445ns ± 4% -36.58% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 604ns ± 1% 311ns ± 1% -48.54% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 179µs ± 3% 188µs ± 2% +5.30% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 26.0 ± 0% 25.0 ± 0% -3.85% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 8.00 ± 0% 7.00 ± 0% -12.50% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 8.49k ± 0% 8.49k ± 0% -0.01% (p=0.000 n=8+8) Change-Id: I6247ac3fd66a63d9acb902cbd192094ee3d151c3 Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/185147 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-06-27 10:54:42 -07:00
type mapInfo struct {
goType reflect.Type
keyWiretag uint64
valWiretag uint64
keyFuncs ifaceCoderFuncs
valFuncs ifaceCoderFuncs
keyZero interface{}
valZero interface{}
newVal func() interface{}
}
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
func encoderFuncsForMap(fd pref.FieldDescriptor, ft reflect.Type) (funcs pointerCoderFuncs) {
// TODO: Consider generating specialized map coders.
keyField := fd.MapKey()
valField := fd.MapValue()
keyWiretag := wire.EncodeTag(1, wireTypes[keyField.Kind()])
valWiretag := wire.EncodeTag(2, wireTypes[valField.Kind()])
keyFuncs := encoderFuncsForValue(keyField, ft.Key())
valFuncs := encoderFuncsForValue(valField, ft.Elem())
internal/impl: add fast-path unmarshal Benchmarks run with: go test ./benchmarks/ -bench=Wire -benchtime=500ms -benchmem -count=8 Fast-path vs. parent commit: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 1.35µs ± 2% 0.45µs ± 4% -67.01% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 1.07µs ± 1% 0.31µs ± 1% -71.04% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 691µs ± 2% 188µs ± 2% -72.78% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 60.0 ± 0% 25.0 ± 0% -58.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 42.0 ± 0% 7.0 ± 0% -83.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 28.6k ± 0% 8.5k ± 0% -70.34% (p=0.000 n=8+8) Fast-path vs. -v1: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 702ns ± 1% 445ns ± 4% -36.58% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 604ns ± 1% 311ns ± 1% -48.54% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 179µs ± 3% 188µs ± 2% +5.30% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 26.0 ± 0% 25.0 ± 0% -3.85% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 8.00 ± 0% 7.00 ± 0% -12.50% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 8.49k ± 0% 8.49k ± 0% -0.01% (p=0.000 n=8+8) Change-Id: I6247ac3fd66a63d9acb902cbd192094ee3d151c3 Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/185147 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-06-27 10:54:42 -07:00
mapi := &mapInfo{
goType: ft,
keyWiretag: keyWiretag,
valWiretag: valWiretag,
keyFuncs: keyFuncs,
valFuncs: valFuncs,
keyZero: reflect.Zero(ft.Key()).Interface(),
valZero: reflect.Zero(ft.Elem()).Interface(),
}
switch valField.Kind() {
case pref.GroupKind, pref.MessageKind:
mapi.newVal = func() interface{} {
return reflect.New(ft.Elem().Elem()).Interface()
}
}
funcs = pointerCoderFuncs{
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
size: func(p pointer, tagsize int, opts marshalOptions) int {
return sizeMap(p, tagsize, ft, keyFuncs, valFuncs, opts)
},
marshal: func(b []byte, p pointer, wiretag uint64, opts marshalOptions) ([]byte, error) {
return appendMap(b, p, wiretag, keyWiretag, valWiretag, ft, keyFuncs, valFuncs, opts)
},
internal/impl: add fast-path unmarshal Benchmarks run with: go test ./benchmarks/ -bench=Wire -benchtime=500ms -benchmem -count=8 Fast-path vs. parent commit: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 1.35µs ± 2% 0.45µs ± 4% -67.01% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 1.07µs ± 1% 0.31µs ± 1% -71.04% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 691µs ± 2% 188µs ± 2% -72.78% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 60.0 ± 0% 25.0 ± 0% -58.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 42.0 ± 0% 7.0 ± 0% -83.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 28.6k ± 0% 8.5k ± 0% -70.34% (p=0.000 n=8+8) Fast-path vs. -v1: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 702ns ± 1% 445ns ± 4% -36.58% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 604ns ± 1% 311ns ± 1% -48.54% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 179µs ± 3% 188µs ± 2% +5.30% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 26.0 ± 0% 25.0 ± 0% -3.85% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 8.00 ± 0% 7.00 ± 0% -12.50% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 8.49k ± 0% 8.49k ± 0% -0.01% (p=0.000 n=8+8) Change-Id: I6247ac3fd66a63d9acb902cbd192094ee3d151c3 Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/185147 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-06-27 10:54:42 -07:00
unmarshal: func(b []byte, p pointer, wtyp wire.Type, opts unmarshalOptions) (int, error) {
return consumeMap(b, p, wtyp, mapi, opts)
},
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
}
if valFuncs.isInit != nil {
funcs.isInit = func(p pointer) error {
return isInitMap(p, ft, valFuncs.isInit)
}
}
return funcs
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
}
const (
mapKeyTagSize = 1 // field 1, tag size 1.
mapValTagSize = 1 // field 2, tag size 2.
)
internal/impl: add fast-path unmarshal Benchmarks run with: go test ./benchmarks/ -bench=Wire -benchtime=500ms -benchmem -count=8 Fast-path vs. parent commit: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 1.35µs ± 2% 0.45µs ± 4% -67.01% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 1.07µs ± 1% 0.31µs ± 1% -71.04% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 691µs ± 2% 188µs ± 2% -72.78% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 60.0 ± 0% 25.0 ± 0% -58.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 42.0 ± 0% 7.0 ± 0% -83.33% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 28.6k ± 0% 8.5k ± 0% -70.34% (p=0.000 n=8+8) Fast-path vs. -v1: name old time/op new time/op delta Wire/Unmarshal/google_message1_proto2-12 702ns ± 1% 445ns ± 4% -36.58% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 604ns ± 1% 311ns ± 1% -48.54% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 179µs ± 3% 188µs ± 2% +5.30% (p=0.000 n=7+8) name old allocs/op new allocs/op delta Wire/Unmarshal/google_message1_proto2-12 26.0 ± 0% 25.0 ± 0% -3.85% (p=0.000 n=8+8) Wire/Unmarshal/google_message1_proto3-12 8.00 ± 0% 7.00 ± 0% -12.50% (p=0.000 n=8+8) Wire/Unmarshal/google_message2-12 8.49k ± 0% 8.49k ± 0% -0.01% (p=0.000 n=8+8) Change-Id: I6247ac3fd66a63d9acb902cbd192094ee3d151c3 Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/185147 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-06-27 10:54:42 -07:00
func consumeMap(b []byte, p pointer, wtyp wire.Type, mapi *mapInfo, opts unmarshalOptions) (int, error) {
mp := p.AsValueOf(mapi.goType)
if mp.Elem().IsNil() {
mp.Elem().Set(reflect.MakeMap(mapi.goType))
}
m := mp.Elem()
if wtyp != wire.BytesType {
return 0, errUnknown
}
b, n := wire.ConsumeBytes(b)
if n < 0 {
return 0, wire.ParseError(n)
}
var (
key = mapi.keyZero
val = mapi.valZero
)
if mapi.newVal != nil {
val = mapi.newVal()
}
for len(b) > 0 {
num, wtyp, n := wire.ConsumeTag(b)
if n < 0 {
return 0, wire.ParseError(n)
}
b = b[n:]
err := errUnknown
switch num {
case 1:
var v interface{}
v, n, err = mapi.keyFuncs.unmarshal(b, key, num, wtyp, opts)
if err != nil {
break
}
key = v
case 2:
var v interface{}
v, n, err = mapi.valFuncs.unmarshal(b, val, num, wtyp, opts)
if err != nil {
break
}
val = v
}
if err == errUnknown {
n = wire.ConsumeFieldValue(num, wtyp, b)
if n < 0 {
return 0, wire.ParseError(n)
}
} else if err != nil {
return 0, err
}
b = b[n:]
}
m.SetMapIndex(reflect.ValueOf(key), reflect.ValueOf(val))
return n, nil
}
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
func sizeMap(p pointer, tagsize int, goType reflect.Type, keyFuncs, valFuncs ifaceCoderFuncs, opts marshalOptions) int {
m := p.AsValueOf(goType).Elem()
n := 0
if m.Len() == 0 {
return 0
}
iter := mapRange(m)
for iter.Next() {
ki := iter.Key().Interface()
vi := iter.Value().Interface()
size := keyFuncs.size(ki, mapKeyTagSize, opts) + valFuncs.size(vi, mapValTagSize, opts)
n += wire.SizeBytes(size) + tagsize
}
return n
}
func appendMap(b []byte, p pointer, wiretag, keyWiretag, valWiretag uint64, goType reflect.Type, keyFuncs, valFuncs ifaceCoderFuncs, opts marshalOptions) ([]byte, error) {
m := p.AsValueOf(goType).Elem()
var err error
if m.Len() == 0 {
return b, nil
}
if opts.Deterministic() {
keys := m.MapKeys()
sort.Sort(mapKeys(keys))
for _, k := range keys {
b, err = appendMapElement(b, k, m.MapIndex(k), wiretag, keyWiretag, valWiretag, keyFuncs, valFuncs, opts)
if err != nil {
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
return b, err
}
}
return b, nil
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
}
iter := mapRange(m)
for iter.Next() {
b, err = appendMapElement(b, iter.Key(), iter.Value(), wiretag, keyWiretag, valWiretag, keyFuncs, valFuncs, opts)
if err != nil {
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
return b, err
}
}
return b, nil
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
}
func appendMapElement(b []byte, key, value reflect.Value, wiretag, keyWiretag, valWiretag uint64, keyFuncs, valFuncs ifaceCoderFuncs, opts marshalOptions) ([]byte, error) {
ki := key.Interface()
vi := value.Interface()
b = wire.AppendVarint(b, wiretag)
size := keyFuncs.size(ki, mapKeyTagSize, opts) + valFuncs.size(vi, mapValTagSize, opts)
b = wire.AppendVarint(b, uint64(size))
b, err := keyFuncs.marshal(b, ki, keyWiretag, opts)
if err != nil {
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
return b, err
}
b, err = valFuncs.marshal(b, vi, valWiretag, opts)
if err != nil {
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
return b, err
}
return b, nil
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
}
func isInitMap(p pointer, goType reflect.Type, isInit func(interface{}) error) error {
m := p.AsValueOf(goType).Elem()
if m.Len() == 0 {
return nil
}
iter := mapRange(m)
for iter.Next() {
if err := isInit(iter.Value().Interface()); err != nil {
return err
}
}
return nil
}
internal/impl: add fast-path marshal implementation This is a port of the v1 table marshaler, with some substantial cleanup and refactoring. Benchstat results from the protobuf reference benchmark data comparing the v1 package with v2, with AllowPartial:true set for the new package. This is not an apples-to-apples comparison, since v1 doesn't have a way to disable required field checks. Required field checks in v2 package currently go through reflection, which performs terribly; my initial experimentation indicates that fast-path required field checks will not add a large amount of cost; these results are incomplete but not wholly inaccurate. name old time/op new time/op delta /dataset.google_message3_1.pb/Marshal-12 219ms ± 1% 232ms ± 1% +5.85% (p=0.004 n=6+5) /dataset.google_message2.pb/Marshal-12 261µs ± 3% 248µs ± 1% -5.14% (p=0.002 n=6+6) /dataset.google_message1_proto2.pb/Marshal-12 681ns ± 2% 637ns ± 3% -6.53% (p=0.002 n=6+6) /dataset.google_message1_proto3.pb/Marshal-12 1.10µs ± 8% 0.99µs ± 3% -9.63% (p=0.002 n=6+6) /dataset.google_message3_3.pb/Marshal-12 44.2ms ± 3% 35.2ms ± 1% -20.28% (p=0.004 n=6+5) /dataset.google_message4.pb/Marshal-12 91.4ms ± 2% 94.9ms ± 2% +3.78% (p=0.002 n=6+6) /dataset.google_message3_2.pb/Marshal-12 78.7ms ± 6% 80.8ms ± 4% ~ (p=0.310 n=6+6) /dataset.google_message3_4.pb/Marshal-12 10.6ms ± 3% 10.6ms ± 8% ~ (p=0.662 n=5+6) /dataset.google_message3_5.pb/Marshal-12 675ms ± 4% 510ms ± 2% -24.40% (p=0.002 n=6+6) /dataset.google_message3_1.pb/Marshal 219ms ± 1% 236ms ± 7% +8.06% (p=0.004 n=5+6) /dataset.google_message2.pb/Marshal 257µs ± 1% 250µs ± 3% ~ (p=0.052 n=5+6) /dataset.google_message1_proto2.pb/Marshal 685ns ± 1% 628ns ± 1% -8.41% (p=0.008 n=5+5) /dataset.google_message1_proto3.pb/Marshal 1.08µs ± 1% 0.98µs ± 2% -9.31% (p=0.004 n=5+6) /dataset.google_message3_3.pb/Marshal 43.7ms ± 1% 35.1ms ± 1% -19.76% (p=0.002 n=6+6) /dataset.google_message4.pb/Marshal 93.4ms ± 4% 94.9ms ± 2% ~ (p=0.180 n=6+6) /dataset.google_message3_2.pb/Marshal 105ms ± 2% 98ms ± 7% -6.81% (p=0.009 n=5+6) /dataset.google_message3_4.pb/Marshal 16.3ms ± 6% 15.7ms ± 3% -3.44% (p=0.041 n=6+6) /dataset.google_message3_5.pb/Marshal 676ms ± 4% 504ms ± 2% -25.50% (p=0.004 n=6+5) Change-Id: I72cc4597117f4cf5d236ef505777d49dd4a5f75d Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/171020 Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>
2019-04-01 13:49:56 -07:00
// mapKeys returns a sort.Interface to be used for sorting the map keys.
// Map fields may have key types of non-float scalars, strings and enums.
func mapKeys(vs []reflect.Value) sort.Interface {
s := mapKeySorter{vs: vs}
// Type specialization per https://developers.google.com/protocol-buffers/docs/proto#maps.
if len(vs) == 0 {
return s
}
switch vs[0].Kind() {
case reflect.Int32, reflect.Int64:
s.less = func(a, b reflect.Value) bool { return a.Int() < b.Int() }
case reflect.Uint32, reflect.Uint64:
s.less = func(a, b reflect.Value) bool { return a.Uint() < b.Uint() }
case reflect.Bool:
s.less = func(a, b reflect.Value) bool { return !a.Bool() && b.Bool() } // false < true
case reflect.String:
s.less = func(a, b reflect.Value) bool { return a.String() < b.String() }
default:
panic(fmt.Sprintf("unsupported map key type: %v", vs[0].Kind()))
}
return s
}
type mapKeySorter struct {
vs []reflect.Value
less func(a, b reflect.Value) bool
}
func (s mapKeySorter) Len() int { return len(s.vs) }
func (s mapKeySorter) Swap(i, j int) { s.vs[i], s.vs[j] = s.vs[j], s.vs[i] }
func (s mapKeySorter) Less(i, j int) bool {
return s.less(s.vs[i], s.vs[j])
}