Asynchronous management of access requests to control power consumption
Systems and methods are disclosed for asynchronous management of access requests to control power consumption. In some cases, by asynchronously managing power within a system, multiple dies of a NVM can simultaneously draw current in order to match the power demand. In particular, an arbiter of the system can receive multiple requests to draw current, where each request may be associated with a different die of the NVM. In some embodiments, the arbiter can determine the servicing order using the time of arrival of the request (e.g., a first-in, first-out scheme). In other embodiments, the arbiter can simultaneously service multiple requests so long as the servicing of the multiple requests does not exceed a power budget.
1. A method for management of power consumption, the method comprising:
receiving a plurality of requests and storing the received requests in a queue, where each request is a request to access a die of a plurality of dies of a non-volatile memory and each request is associated with a received time of arrival, wherein the plurality of request comprises a first request associated with a first die and a first arrival time, a second request associated with the first die and a second arrival time, and a third request associated with a second die and a third arrival time;
selecting a subset of the requests from the queue of requests, wherein each request in the subset is associated with a different die, and wherein the subset comprises the first and third requests;
determining whether adding a current draw of the subset of requests to a current consumption will exceed a power budget;
in response to determining that adding the current draw of the subset of requests to the current consumption will not exceed the power budget, at a first time period, asynchronously servicing the first and third requests by enabling the dies associated with first and third requests to concurrently draw power, whereby the asynchronous servicing of requests is evidenced by servicing the third request prior to servicing the second request even though the second request was received before the third request; and
at a second time period immediately following the first time period, asynchronously servicing the second request if it is determined that a current draw of the second request will not exceed the power budget by enabling the die associated with the second request to draw power.
2. The method of claim 1 , further comprising updating the current consumption with the current draw of the subset of requests.
3. The method of claim 2 wherein the updating comprises adding the current draw of the subset of request to the current consumption.
4. The method of claim 1 , further comprising enabling the die of the non-volatile memory to draw current from a power source.
5. The method of claim 1 , further comprising in response to determining that adding the current draw of the request to the current consumption will exceed the power budget, delaying the servicing of the request.
6. The method of claim 1 , further comprising:
determining whether at least one of the requests has completed; and
in response to determining that the request has completed, updating the current consumption with the current draw of the at least one of the requests.
7. The method of claim 6 , wherein the updating further comprises subtracting the current draw of the at least one of the requests from the current consumption.
8. The method of claim 6 , further comprising removing the at least one of the requests from the queue of requests.
9. The method of claim 1 , further comprising maintaining the queue in an arbiter.
10. The method of claim 9 , wherein the arbiter is operative to asynchronously enable the dies associated with the requests to draw power from a power source.