Alex

Currently, it is difficult to say that observed anomalies are fully unrelated to oneTBB library. The deadlocks are mostly related to the testing approach but the assertion seems to be...

@phprus , can you please check that test does not fail in debug? Reopening the issue because it is not fixed in release yet

Thank you for the confirmation. As for release, we have some experiments in [dev/alexey-katranov/exp-seq-cst](https://github.com/oneapi-src/oneTBB/tree/dev/alexey-katranov/exp-seq-cst) but it is not clear if we go with such approach because it might affect performance.

> `test_global_control` and `test_arena_constraints` (https://github.com/oneapi-src/oneTBB/issues/756) works without hangs. Sounds great. Unfortunately, `conformance_parallel_pipeline` still hangs after ~859800 runs...

@ekovanova, can you please look at the patch?

> The objects T that the shared pointers of the vector are pointing to have to be accessed atomically in our use case. I believe you have data race here...

> I believe it is still possible to create such a thread-safe std::optional-like class, which will not be in that gray area We failed to found simple optional-like solution in...

It seems we have two issues: 1. Broken serialization of worker thread requests (more threads than system concurrency are created) 2. Broken barrier lifetime (some threads from previous epochs are...

> M1 Max has 10 cores (2 efficient and 8 performance cores) Thank you for the info, I was thinking about 8 cores, so the assumption above is not correct.

It seems the testing approach is broken on ARM, any test using `utils::SpinBarrier` might hang without real issue. We are thinking about better approach for testing.