Wwise SDK 2021.1.14

Interface for batched deferred low-level I/O transfers. Used by streaming devices created with the AK_SCHEDULER_DEFERRED_LINED_UP flag. This I/O transfer handshaking method is preferred when you want to hook I/O to your own I/O streaming technology, and you want to submit multiple I/O requests in one call, so as to allow for better opportunities for CPU and I/O performance. You may queue them into your own system, and even use the heuristics passed down to this level for your convenience. Note that the requests are always sent in the order that the Stream Manager considers to be the most appropriate. You may receive less than AkDeviceSettings::uMaxConcurrentIO at any given time. The number of concurrent transfers depends on the number of streams running in the high-level streaming device, and on its target buffering length and granularity. Your advantage at this level is to be aware of file placement, so you may try to re-order requests in order to minimize seeking on disk. Calls to BatchRead()/BatchWrite() should return as soon as possible. You need to call AkAsyncIOTransferInfo::pCallback for an individual item, or AkBatchIOCallback for a collection of items as soon as a transfer is completed. Cancel() is provided in order to inform you that the streaming device will flush this transfer upon completion. You may implement it or not. In all cases, you must call the callback.

AkStreamMgrModule.h の 305 行目に定義があります。