membuffer: implement cache for ART

[you06]

ref pingcap/tidb#55287

ART can also benefit from the cache of last visited leaf, this PR implement the cache, it's almost same as RBT's implementation.

The test result of new added BenchmarkMemBufferCache:

// RBT, with cache, 1 cache-miss and 10 cache-hits for each key
BenchmarkMemBufferCache/RBT
BenchmarkMemBufferCache/RBT-32          10000000               185.9 ns/op             0 B/op          0 allocs/op

// master, no cache, 11 cache-miss for each key
BenchmarkMemBufferCache/ART
BenchmarkMemBufferCache/ART-32          10000000               223.3 ns/op             0 B/op          0 allocs/op

// this PR, with cache, 1 cache-miss and 10 cache-hits for each key
BenchmarkMemBufferCache/ART
BenchmarkMemBufferCache/ART-32          10000000               126.6 ns/op             0 B/op          0 allocs/op

[ekexium]

A benchmark result of its effect in TiDB's use case would be more convincing.

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: cfzjywxk, ekexium

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [cfzjywxk,ekexium]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[ti-chi-bot]

[LGTM Timeline notifier]

Timeline:

• 2024-10-08 08:38:54.519176158 +0000 UTC m=+948889.939389154: ☑️ agreed by ekexium.
• 2024-10-08 08:53:04.921955496 +0000 UTC m=+949740.342168508: ☑️ agreed by cfzjywxk.

[you06]

A benchmark result of its effect in TiDB's use case would be more convincing.

Little improvement, see TiDB's PR.