In what way is this different from the wait-free circular buffer implementation described by Sape Mullender [1] (of Plan 9 fame). I listened in to a presentation he gave last november and one of the take-aways was that it scales poorly with the number of cores.
He implemented this in Plan 9, but the performance gain was not tremendous due to shared memory effects among the cores.
He implemented this in Plan 9, but the performance gain was not tremendous due to shared memory effects among the cores.
[1] http://lore.ua.ac.be/Teaching/CapitaMaster/antwerpen-threads