Memory issue with `sorting_with_executors`

[alexbarev]

1. Run sorting_with_executors benchmark using the std::execution::par_unseq policy.

2. Memory consumption quickly exceeds my machine's availability of 60GB after the second test variant finishes: sorting_with_executors/par_unseq/4194304/

Observations:

1. Memory does not decrease between tests with inputs of different sizes.

2. I tried moving the policy inside the loop, but it did not resolve the issue:

   for (auto _ : state) {
       auto local_policy = std::execution::par_unseq; // Create policy here
       std::reverse(local_policy, array.begin(), array.end());
       std::sort(local_policy, array.begin(), array.end());
       bm::DoNotOptimize(array.size());
       bm::DoNotOptimize(local_policy); // even tried this just in case
   }
   

3. I noticed that only 16 threads are spawned, which matches the number of available CPU's. The issue is puzzling, and I am yet to understand its root cause.