Fair Change Detection Benchmarking

[re0312]

Objective

• [Merged by Bors] - Change Detection Benchmarks #4972 introduce a benchmark to measure chang detection performance
• However,it uses iter_batch cause a lot of overhead in clone data to each routine closure(it feels like a bug initer_batch ) and constructs new query in every iter.This overhead masks the real change detection throughput we want to measure. Instead of evaluating raw change detection, the benchmark ends up dominated by data cloning and allocation costs.

Solution

• Use iter_batch_ref to reduce the benchmark overload
• Use cached query to better reflect real-world usage scenarios.
• Add more benmark

---

Changelog

[james7132]

This looks well motivated and the reasoning works out, but I'd like to see how much of an impact this has on the final benchmark numbers. This isn't to say the PR will be blocked on them changing in a particular direction, but it's best we show what impact this has.

[re0312]

This looks well motivated and the reasoning works out, but I'd like to see how much of an impact this has on the final benchmark numbers. This isn't to say the PR will be blocked on them changing in a particular direction, but it's best we show what impact this has.

I do a benchmark in main and this pr,results are as follows

main (entities from 50000 to 60000 ,the detection time increase 2x~3x,this appears to be an unreasonable outcome.)

pr(added the 60000 entities benchmark to compare)

time increase linear with entity count