std::execution::unseq 启用了哪些优化？

Question

In C++20, another execution policy: std::execution::unseq was added. It means, that an algorithm will be executed on current thread, but without guarantee that operations will be done in the order of the elements.

What is the rationale for adding that to the language, what optimizations does it enable? As far as I looked through implementation, this flag is pretty much treated exactly as seq , unless I missed something (what is likely). In my experience, compilers already recognize loops that they can vectorize (for trivial types) and vectorization is possible there without any special flags due to "as-if" rule. So, what does std::execution::unseq change?

Answer 1

So, what does std::execution::unseq change?

Standard says (latest draft):

[execpol.unseq]

The class unsequenced_policy is an execution policy type used as a unique type to disambiguate parallel algorithm overloading and indicate that a parallel algorithm's execution may be vectorized , eg, executed on a single thread using instructions that operate on multiple data items.

---

[algorithms.parallel.exec]

The invocations of element access functions in parallel algorithms invoked with an execution policy object of type execution​::​unsequenced_policy are permitted to execute in an unordered fashion in the calling thread of execution, unsequenced with respect to one another in the calling thread of execution.

[Note 4: This means that multiple function object invocations can be interleaved on a single thread of execution, which overrides the usual guarantee from [intro.execution] that function executions do not overlap with one another. — end note]

---

compilers already recognize loops that they can vectorize

Some compilers may sometimes be smart enough to detect that vectorisation is possible, and guess that it is beneficial. But it isn't simple and won't always work. By telling the compiler that vectorisation is OK and desirable, there won't be guessing involved.