Investigate possible histogram performance improvements

[mic-max]

• Put a cap on the histogram explicit bound count? 10 million worked fine https://github.com/open-telemetry/opentelemetry-dotnet/pull/3252#discussion_r912271977
• Can we avoid saving a copy of the histogram bounds array when in binary search mode (currently enabled when bound count is 50+)
• Store the BST in the AggregatorStore and pass that to the MetricPoint constructor https://github.com/open-telemetry/opentelemetry-dotnet/pull/3252#discussion_r912219474
• Avoid using a delegate function for switching between linear and binary search https://github.com/open-telemetry/opentelemetry-dotnet/pull/3252#discussion_r917166136
• Radix sort suggested by @reyang
• Use SIMD vectorization to find histogram bucket faster https://devblogs.microsoft.com/dotnet/performance_improvements_in_net_7/#vectorization
• Dedicated methods to avoid scenario specific if-checks potential perf increase https://github.com/open-telemetry/opentelemetry-dotnet/pull/2735#discussion_r971482389

[github-actions[bot]]

This issue was marked stale due to lack of activity and will be closed in 7 days. Commenting will instruct the bot to automatically remove the label. This bot runs once per day.

[github-actions[bot]]

Closed as inactive. Feel free to reopen if this issue is still a concern.