Systems and methods for exploiting queues and transitional storage for improved low-latency high-bandwidth on-die data retrieval
Apparatuses including general-purpose graphics processing units and graphics multiprocessors that exploit queues or transitional buffers for improved low-latency high-bandwidth on-die data retrieval are disclosed. In one embodiment, a graphics multiprocessor includes at least one compute engine to provide a request, a queue or transitional buffer, and logic coupled to the queue or transitional buffer. The logic is configured to cause a request to be transferred to a queue or transitional buffer for temporary storage without processing the request and to determine whether the queue or transitional buffer has a predetermined amount of storage capacity.
1. A graphics multiprocessor, comprising:
at least one processing engine to provide a request;
a queue or transitional buffer to function as cache memory and to provide on-die or on chip data retrieval for the graphics multiprocessor; and
logic on die or on chip with the queue or transitional buffer, the logic is configured to cause the request of the at least one processing engine to be transferred to the queue or transitional buffer for temporary storage instead of transferring the request to off chip memory when the queue or transitional buffer has a predetermined amount of storage capacity.
2. The graphics multiprocessor of claim 1 , wherein the logic is further configured to determine whether the queue or transitional buffer has a predetermined amount of storage capacity including greater than a threshold level of storage capacity.
3. The graphics multiprocessor of claim 2 , wherein the logic is further configured to temporarily store the request in the queue or transitional buffer without processing the request when the queue or transitional buffer has the predetermined amount of storage capacity.
4. The graphics multiprocessor of claim 3 , wherein the logic is further configured to transfer the request to a different queue, buffer, or a next level of memory when the queue or transitional buffer lacks the predetermined amount of storage capacity.
5. The graphics multiprocessor of claim 1 , wherein the logic comprises a cache controller or memory interface controller.
6. The graphics multiprocessor of claim 1 , wherein the queue or transitional buffer is located in any location on die or on chip with the logic and the at least one processing engine.
7. The graphics multiprocessor of claim 1 , wherein the queue or transitional buffer provides functionality of cache memory.
8. A graphics processing unit, comprising:
at least one processing engine to provide a request;
a queue or transitional buffer to function as cache memory and to provide on-die or on chip data retrieval for the graphics processing unit; and
a controller coupled to and on die or on chip with the queue or transitional buffer and the at least one processing engine, the controller is configured to cause the request of the at least one processing engine to be transferred to the queue or transitional buffer for temporary storage instead of transferring the request to off chip memory when the queue or transitional buffer has a predetermined amount of storage capacity.
9. The graphics processing unit of claim 8 , wherein the controller is further configured to determine whether the queue or transitional buffer has a predetermined amount of storage capacity including greater than a threshold level of storage capacity.
10. The graphics processing unit of claim 9 , wherein the controller is further configured to temporarily stored the request in the queue or transitional buffer without processing the request when the queue or transitional buffer has the predetermined amount of storage capacity.
11. The graphics processing unit of claim 8 , wherein the controller is further configured to transfer the request to a different queue, buffer, or a next level of memory when the queue or transitional buffer lacks the predetermined amount of storage capacity.
12. The graphics processing unit of claim 8 , wherein the controller comprises a cache controller or memory interface controller.
13. The graphics processing unit of claim 8 , wherein the queue or transitional buffer is located in any location on die or on.
14. The graphics processing unit of claim 8 , wherein the queue or transitional buffer provides functionality of cache memory.
15. A computer implemented method, comprising:
receiving with a queue or transitional buffer having functionality of cache memory, a request from at least one processing engine of a graphics multiprocessor;
determining, with a controller, whether the queue or transitional buffer has a predetermined amount of storage capacity; and
temporarily storing the request in the queue or transitional buffer instead of transferring the request to off chip memory when the queue or transitional buffer has the predetermined amount of storage capacity, wherein the queue or transitional buffer is located in any location on die or on chip with the controller and the at least one processing engine of the graphics multiprocessor.
16. The computer implemented method of claim 15 , further comprising:
determining whether the queue or transitional buffer has greater than a threshold level of storage capacity.
17. The computer implemented method of claim 15 , further comprising:
temporarily storing the request in the queue or transitional buffer without processing the request when the queue or transitional buffer has the predetermined amount of storage capacity.
18. The computer implemented method of claim 15 , further comprising:
transferring the request to a different queue, buffer, or a next level of memory when the queue or transitional buffer lacks the predetermined amount of storage capacity.
19. The computer implemented method of claim 15 , wherein the queue or transitional buffer is located in any location on die or on chip.
20. The computer implemented method of claim 15 , wherein the queue or transitional buffer provides functionality of cache memory.