omp 和 bthread之间会有竞争，导致有一些长尾

目前我们服务用的bthread，服务依赖的 库用的omp做计算，目前看在不调整参数的情况下，omp 和 bthread之间会有竞争，导致有一些长尾的问题，目前调整了bthread的数量和omp使用的数量，看着整体有缓解。
想问下目前这方面有啥比较通用可行的解决方案吗。

The issue names no source files, tests, or entry points. Start by reproducing the long-tail latency while varying bthread and OpenMP thread counts, then inspect the relevant bthread and OpenMP scheduling paths. Done means identifying a generally applicable mitigation and documenting its trade-offs and validation results.

Written by the indexing model from the issue text.