Thanks for the feedback. As for the last getAcquire, I think you're right. Considering the prior loadLoadFence, it can be changed to getOpaque.

Read operations will scale only with increased CPU core count. As for writes, they might scale better in Go with increased number of goroutines as long as the hash table bucket count is large enough. That's because of cheaper context switches in green threads vs OS threads. Project Loom may indeed change that.

memory_order_relaxed is coherent and opaque is not. Your link demonstrates that getOpaque() is coherent (writes on the same variable are guaranteed to be observed in a total order) even on non-MCA architectures. Am I missing something here?

Yes, monotonicity is not guaranteed on such counter. The counter is nothing more than an excuse to compare different semantics and operations. As for LongAdder, it's mentioned in the post along with pros and cons. Thanks for the comment on relaxed vs. Opaque. That's interesting.