| 摘要 |
A method and apparatus for allocation and control of memory bandwidth within a video graphics system is accomplished by first determining the memory bandwidth needs of each of the plurality of memory clients in the video graphics system. Based on this determination, a plurality of timers are configured, wherein each of the timers corresponds to one of the plurality of memory clients. The timers associated with the memory clients store two values. One value indicates the memory access interval for the corresponding client, which determines the spacing between memory access requests that can be issued by that particular client. The other value stored in the time is a memory access limit value, which determines the maximum length of a protected access to the memory by that particular client. A memory controller in the system receives requests from the plurality of clients and determines the priority of the different requests. The memory controller grants the requests with the highest priority, which may result in the termination of a current memory access. However, if the current memory access has equal or greater priority than a received access, the current memory access will not be terminated until the time associated with its memory access limit has expired.
|
