BENCHMARKS.md

Benchmarks

Summary

• Slice with preallocation wins in most benchmarks.
• Preallocating improves aggregate results in all cases, so use it if you can.

Data

goos: darwin goarch: amd64 pkg: github.com/fgm/container cpu: Intel(R) Core(TM) i7-4980HQ CPU @ 2.80GHz

Storage	Queue.Enqueue	Queue.Dequeue	Stack.Push	Stack.Pop
Slice prealloc	9 ns/op	10 ns/op	8 ns/op	7 ns/op
Slice raw	25 ns/op	11 ns/op	20 ns/op	5 ns/op
List	144 ns/op	10 ns/op	130 ns/op	56 ns/op
ListSP prealloc	533 ns/op	52 ns/op	96 ns/op	18 ns/op
ListSP raw	368 ns/op	56 ns/op	385 ns/op	61 ns/op
ListIP prealloc	17 ns/op	14 ns/op	11 ns/op	11 ns/op
ListIP raw	139 ns/op	84 ns/op	121 ns/op	9 ns/op

Ranking per op

The >> operator indicates a big decrease in performance.

- Enqueue: SP > IPP > SR              >> IPR >  L  > SPR > SPP
- Dequeue: SP >  L  > SR  > IPP       >> SPP > SPR > IPR
- Push:    SP > IPP > SR              >> SPP > IPR >  L  > SPR
- Pop:     SR >  SP > IPR > IPP > SPP >>  L  > SPR

Scores

Scores calculated by adding the ranking of implementations for each operation. Less is better.

Implementation	Enqueue	Dequeue	Push	Pop	Total
Slice, prealloc	1	1	1	2	5
Slice, raw	3	3	3	1	10
Internal pool, prealloc	2	4	2	4	12
List	5	2	6	6	19
Internal pool, raw	4	7	5	3	19
Sync pool, prealloc	7	5	5	5	22
Sync pool, raw	6	6	7	6	25

Although the list with internal pool comes close, it never outperforms the simple slice-based implementation.