Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/rsheremeta/gob-proto-bench
Benchmarking (de)serialization of Gob vs Protobuf
https://github.com/rsheremeta/gob-proto-bench
benchmark benchmarking go gob golang proto protobuf protobuf3 protocol-buffers serialization
Last synced: 9 days ago
JSON representation
Benchmarking (de)serialization of Gob vs Protobuf
- Host: GitHub
- URL: https://github.com/rsheremeta/gob-proto-bench
- Owner: RSheremeta
- Created: 2023-06-07T08:46:44.000Z (over 1 year ago)
- Default Branch: master
- Last Pushed: 2023-06-13T15:04:28.000Z (over 1 year ago)
- Last Synced: 2024-01-29T18:14:15.287Z (10 months ago)
- Topics: benchmark, benchmarking, go, gob, golang, proto, protobuf, protobuf3, protocol-buffers, serialization
- Language: Go
- Homepage:
- Size: 24.4 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# gob-proto-bench
#### Benchmarking Gob vs Protobuf.
This is Part 2 of the experiment of measuring Gob's (de)serialization performance. See Part 1 [here](https://github.com/RSheremeta/gob-serialization).
This repo is made just out of curiosity and has an experimental reasoning.
### Description
There are 4 structs: `Tiny`, `Medium`, `Big`, `Huge` and a struct called `ComplexAndHuge` which holds `type complexMap map[string]*Huge`.
Each of them has a respective size and complexity.
And each of them builds the following targets to run encoding & decoding benchmarks against:
- `Single` means a single struct - eg - `Tiny{}`
- `Slice` means a slice of given struct - eg - `[]*Tiny`
### Usage
You can just look at and analyze the results I got while running on my machine by observing csv samples in **results_run/sample/** dir.
Also, feel free to play on your own by doing the following:
Run a corresponding **Make** command in terminal and observe the results in **results_run/** dir.
Otherwise run `make run-all` in order to run everything. Warning: beware as it's a very CPU/RAM consuming operation.
### Results
**Software and Hardware:**
The experiment is run on:
- MacBook Pro 2021 16" with macOS Ventura 13.2.1
- 32gb RAM
- Apple M1 Pro CPU
- Go 1.20.3
Here is the full analysis of mine about all the results run: [Analysis.md](https://github.com/RSheremeta/gob-proto-bench/blob/master/Analysis.md)
[**map_slice_decode.csv:**](https://github.com/RSheremeta/gob-proto-bench/blob/master/results_run/sample/map_slice_decode.csv)
```bash
goos: darwin
goarch: arm64
pkg: github.com/RSheremeta/gob-proto-bench/test
BenchmarkDecodeSliceComplexMap/type=GOB_struct_size=huge_complex_map_slice-10 10 573950896 ns/op 633277254 B/op 10645891 allocs/op
BenchmarkDecodeSliceComplexMap/type=Proto_struct_size=huge_complex_map_slice-10 10 2153558850 ns/op 2337695762 B/op 46637425 allocs/op
PASS
ok github.com/RSheremeta/gob-proto-bench/test 35.524s
```
[**all.csv:**](https://github.com/RSheremeta/gob-proto-bench/blob/master/results_run/sample/all.csv)
```bash
goos: darwin
goarch: arm64
pkg: github.com/RSheremeta/gob-proto-bench/test
BenchmarkEncodeSingleComplexMap/type=GOB_struct_size=huge_complex_map-10 10 51486367 ns/op 84000476 B/op 731909 allocs/op
BenchmarkEncodeSingleComplexMap/type=Proto_struct_size=huge_complex_map-10 10 42282621 ns/op 43574969 B/op 466007 allocs/op
BenchmarkDecodeSingleComplexMap/type=GOB_struct_size=huge_complex_map-10 10 55438804 ns/op 36013430 B/op 1064916 allocs/op
BenchmarkDecodeSingleComplexMap/type=Proto_struct_size=huge_complex_map-10 10 43338646 ns/op 46754095 B/op 932749 allocs/op
BenchmarkEncodeSliceComplexMap/type=GOB_struct_size=huge_complex_map_slice-10 10 478913167 ns/op 809327492 B/op 7318093 allocs/op
BenchmarkEncodeSliceComplexMap/type=Proto_struct_size=huge_complex_map_slice-10 10 2021125029 ns/op 2181639455 B/op 23299823 allocs/op
BenchmarkDecodeSliceComplexMap/type=GOB_struct_size=huge_complex_map_slice-10 10 565821717 ns/op 633277498 B/op 10645893 allocs/op
BenchmarkDecodeSliceComplexMap/type=Proto_struct_size=huge_complex_map_slice-10 10 2124560267 ns/op 2337694348 B/op 46637417 allocs/op
BenchmarkEncodeSingle/type=GOB_struct_size=tiny-10 10 2708 ns/op 1136 B/op 20 allocs/op
BenchmarkEncodeSingle/type=GOB_struct_size=medium-10 10 4771 ns/op 1720 B/op 38 allocs/op
BenchmarkEncodeSingle/type=GOB_struct_size=big-10 10 6317 ns/op 2792 B/op 61 allocs/op
BenchmarkEncodeSingle/type=GOB_struct_size=huge-10 10 740108 ns/op 1184470 B/op 5407 allocs/op
BenchmarkEncodeSingle/type=Proto_struct_size=tiny-10 10 950.0 ns/op 16 B/op 1 allocs/op
BenchmarkEncodeSingle/type=Proto_struct_size=medium-10 10 1017 ns/op 48 B/op 1 allocs/op
BenchmarkEncodeSingle/type=Proto_struct_size=big-10 10 1258 ns/op 128 B/op 1 allocs/op
BenchmarkEncodeSingle/type=Proto_struct_size=huge-10 10 464279 ns/op 172032 B/op 1 allocs/op
BenchmarkDecodeSingle/type=GOB_struct_size=tiny-10 10 11879 ns/op 6776 B/op 180 allocs/op
BenchmarkDecodeSingle/type=GOB_struct_size=medium-10 10 15504 ns/op 8728 B/op 237 allocs/op
BenchmarkDecodeSingle/type=GOB_struct_size=big-10 10 18975 ns/op 11379 B/op 317 allocs/op
BenchmarkDecodeSingle/type=GOB_struct_size=huge-10 10 1167321 ns/op 732968 B/op 21628 allocs/op
BenchmarkDecodeSingle/type=Proto_struct_size=tiny-10 10 204.2 ns/op 72 B/op 2 allocs/op
BenchmarkDecodeSingle/type=Proto_struct_size=medium-10 10 675.0 ns/op 256 B/op 6 allocs/op
BenchmarkDecodeSingle/type=Proto_struct_size=big-10 10 925.0 ns/op 680 B/op 14 allocs/op
BenchmarkDecodeSingle/type=Proto_struct_size=huge-10 10 906150 ns/op 934972 B/op 18652 allocs/op
BenchmarkEncodeSlice/type=GOB_struct_size=tiny-10 10 4858 ns/op 1664 B/op 25 allocs/op
BenchmarkEncodeSlice/type=GOB_struct_size=medium-10 10 7342 ns/op 4504 B/op 54 allocs/op
BenchmarkEncodeSlice/type=GOB_struct_size=big-10 10 34996 ns/op 48297 B/op 272 allocs/op
BenchmarkEncodeSlice/type=GOB_struct_size=huge-10 10 34902979 ns/op 65902422 B/op 266108 allocs/op
BenchmarkEncodeSlice/type=Proto_struct_size=tiny-10 10 1329 ns/op 80 B/op 1 allocs/op
BenchmarkEncodeSlice/type=Proto_struct_size=medium-10 10 2862 ns/op 512 B/op 1 allocs/op
BenchmarkEncodeSlice/type=Proto_struct_size=big-10 10 19642 ns/op 6528 B/op 1 allocs/op
BenchmarkEncodeSlice/type=Proto_struct_size=huge-10 10 22906321 ns/op 8314889 B/op 1 allocs/op
BenchmarkDecodeSlice/type=GOB_struct_size=tiny-10 10 12954 ns/op 8035 B/op 219 allocs/op
BenchmarkDecodeSlice/type=GOB_struct_size=medium-10 10 19488 ns/op 11513 B/op 322 allocs/op
BenchmarkDecodeSlice/type=GOB_struct_size=big-10 10 66958 ns/op 40086 B/op 1133 allocs/op
BenchmarkDecodeSlice/type=GOB_struct_size=huge-10 10 54951358 ns/op 36011614 B/op 1064871 allocs/op
BenchmarkDecodeSlice/type=Proto_struct_size=tiny-10 10 1104 ns/op 544 B/op 15 allocs/op
BenchmarkDecodeSlice/type=Proto_struct_size=medium-10 10 3179 ns/op 2872 B/op 66 allocs/op
BenchmarkDecodeSlice/type=Proto_struct_size=big-10 10 34592 ns/op 35080 B/op 708 allocs/op
BenchmarkDecodeSlice/type=Proto_struct_size=huge-10 10 43635229 ns/op 46749566 B/op 932608 allocs/op
PASS
ok github.com/RSheremeta/gob-proto-bench/test 67.650s
```