maybe your test dataset range is too small.

[Ruth-Seven]

go test   . -bench=. -run=NOTEST  -v -cpuprofile=cpu.out -benchtime=1000000x
goos: darwin
goarch: amd64
pkg: github.com/skiplist
cpu: Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
BenchmarkSkipListInsert
BenchmarkSkipListInsert-12                       1000000               224.7 ns/op
BenchmarkSkipListFind
BenchmarkSkipListFind-12                         1000000               151.9 ns/op
BenchmarkSkipListFindBiggerOrEqual
BenchmarkSkipListFindBiggerOrEqual-12            1000000               140.7 ns/op
BenchmarkSkipListDelete
BenchmarkSkipListDelete-12                       1000000                72.63 ns/op

go test   . -bench=. -run=NOTEST  -v -cpuprofile=cpu.out -benchtime=10000000x 
goos: darwin
goarch: amd64
pkg: github.com/skiplist
cpu: Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
BenchmarkSkipListInsert
BenchmarkSkipListInsert-12                      10000000               224.0 ns/op
BenchmarkSkipListFind
BenchmarkSkipListFind-12                        10000000               163.9 ns/op
BenchmarkSkipListFindBiggerOrEqual
BenchmarkSkipListFindBiggerOrEqual-12           10000000               156.8 ns/op
BenchmarkSkipListDelete
BenchmarkSkipListDelete-12                      10000000                80.16 ns/op


go test   . -bench=. -run=NOTEST  -v -cpuprofile=cpu.out -benchtime=100000000x 
goos: darwin
goarch: amd64
pkg: github.com/skiplist
cpu: Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
BenchmarkSkipListInsert
BenchmarkSkipListInsert-12                      100000000              329.4 ns/op
BenchmarkSkipListFind
BenchmarkSkipListFind-12                        100000000              369.1 ns/op
BenchmarkSkipListFindBiggerOrEqual
BenchmarkSkipListFindBiggerOrEqual-12           100000000              318.8 ns/op
BenchmarkSkipListDelete
BenchmarkSkipListDelete-12                      100000000              269.6 ns/op

relative code

[andersfylling]

It's documented in code that it's optimized for up till 34m entries. It would be very interesting to see how it scales compared to other algs, and how increasing the maxLevel constant respectively affects scaling.

skiplist/skiplist.go

Line 40 in 77f5c8d

// efficient for up to 34m entries. If there is a need for much more, please adjust this constant accordingly.