Two level cache memory architecture
View Patent ↗A memory architecture for use in a graphics processor including a main memory, a level one (L1) cache and a level two (L2) cache, coupled between the main memory and the L1 cache is disclosed. The L2 cache stores overlapping requests to the main memory before the requested information is stored in the L1 cache. In this manner, overlapping requests for previously stored information is retrieved from the faster L2 cache as opposed to the relatively slower main memory.
1. A memory architecture comprising:
a plurality of level one caches each comprising texel information and each associated with a corresponding graphics pipeline of a plurality of graphics pipelines;
a common level two cache, operatively coupled to each of the plurality of level one caches, that comprises overlapping fetched texel information resulting from execution of previous memory fetch instructions, wherein the previous memory fetch instructions resulted in storage of requested texel information in at least one corresponding level one cache of the plurality of level one caches;
wherein when a particular level one cache of the plurality of level one caches does not comprise overlapping fetched texel information requested by a subsequent memory fetch instruction, the level two cache is operative to transmit the overlapping fetched texel information requested by the subsequent memory fetch instruction to the particular level one cache of the plurality of level one caches; and
wherein the subsequent memory fetch instruction results in storage of requested texel information at least in the particular level one cache of the plurality of level one caches.
2. The memory architecture of claim 1 , wherein:
each level one cache of the plurality of level one caches comprises a plurality of texture cache blocks; and
wherein one of the plurality of texture cache blocks is operative to receive the subsequent texel fetch instruction.
3. The memory architecture of claim 1 , further comprising:
a main memory operatively coupled to the level two cache; and
wherein when the particular level one cache of the plurality of level one caches and the level two cache do not comprise texel information requested by a second subsequent memory fetch instruction, the main memory is operative to transmit the texel information requested by the second subsequent memory fetch instruction to the level two cache for storage.
4. The memory architecture of claim 3 , wherein the level two cache transmits the texel information requested by the second subsequent texel fetch instruction to the particular level one cache of the plurality of level one caches for storage.
5. A graphics processing device, comprising:
a graphics controller operative to execute memory fetch instructions;
a main memory;
a plurality of level one caches each coupled to the graphics controller, each comprising texel information, and each associated with a corresponding graphics pipeline of a plurality of graphics pipelines; and
a level two cache coupled between the main memory and the plurality of level one caches, the level two cache comprising overlapping fetched texel information resulting from execution of previous memory fetch instructions, wherein the previous memory fetch instructions resulted in storage of requested texel information in at least one corresponding level one cache of the plurality of level one caches;
wherein when a particular level one cache of the plurality of level one caches does not comprise overlapping fetched texel information requested by a subsequent memory fetch instruction, the level two cache transmits the overlapping fetched texel information requested by the subsequent memory fetch instruction to the particular level one cache of the plurality of level one caches; and
wherein subsequent memory fetch instruction results in storage of requested texel information at least in the particular level one cache of the plurality of level one caches.
6. The graphics processing device of claim 5 , wherein the graphics controller is operative to request the subsequent memory fetch instruction.
7. The graphics processing device of claim 6 , wherein the graphics controller comprises a plurality of fetch blocks, wherein one of the plurality of fetch blocks is operative to request the subsequent memory fetch instruction.
8. The graphics processing device of claim 6 , wherein the particular level one cache of the plurality of level one caches transmits the overlapping fetched texel information requested by the subsequent memory fetch instruction to the graphics controller.
9. The graphics processing device of claim 5 , wherein
each of the plurality of level one caches comprises a plurality of texture cache blocks; and
wherein one of the plurality of texture cache blocks is operative to receive the subsequent texel fetch instruction.
10. The graphics processing device of claim 5 , wherein when the particular level one cache of the plurality of level one caches and the level two cache do not comprise texel information requested by a second subsequent memory fetch instruction, the main memory is operative to transmit the texel information requested by the second subsequent memory fetch instruction to the level two cache for storage.
11. The graphics processing device of claim 10 , wherein the level two cache transmits the texel information requested by the second subsequent texel fetch instruction to the particular level one cache of the plurality of level one caches for storage.