Does bar.sync Emit Semaphores Alongside bar.arrive?

[ziyuhuang123]

Does bar.sync also emit semaphores? What I mean is, for example, if I set bar.arrive 0, 128 (executed by 32 threads) and then bar.sync 0, 128 (executed by 96 threads), the total becomes 32 + 96 = 128. This satisfies the condition for sync and should execute normally, correct? (This idea comes from FA3, where the semaphore value equals the sum of arrive and sync calls, not just the arrive count alone.)

[tridao]

You can read the PTX docs