US 12,332,796 B2
Dynamic cache coherence protocol based on runtime interconnect utilization
Keqiang Wu, Palatine, IL (US); Lingxiang Xiang, San Jose, CA (US); Heidi Pan, Burlingame, CA (US); Christopher J. Hughes, Santa Clara, CA (US); and Zhe Wang, San Jose, CA (US)
Assigned to Intel Corporation, Santa Clara, CA (US)
Appl. No. 18/562,743
Filed by Intel Corporation, Santa Clara, CA (US)
PCT Filed Dec. 15, 2021, PCT No. PCT/CN2021/138342
§ 371(c)(1), (2) Date Nov. 20, 2023,
PCT Pub. No. WO2023/108480, PCT Pub. Date Jun. 22, 2023.
Prior Publication US 2024/0303195 A1, Sep. 12, 2024
Int. Cl. G06F 12/0831 (2016.01); G06F 12/084 (2016.01); G06F 12/0891 (2016.01)
CPC G06F 12/0835 (2013.01) [G06F 12/084 (2013.01); G06F 12/0891 (2013.01)] 25 Claims
OG exemplary drawing
 
1. A processor, comprising:
a first cache;
interconnect circuitry to communicate with a second processor via a processor interconnect, wherein the second processor comprises a second cache;
processing circuitry to generate a memory read request for a corresponding memory address of a memory; and
cache controller circuitry to:
detect, based on the memory read request, a cache miss in the first cache, wherein the cache miss indicates that the first cache does not contain a valid copy of data for the corresponding memory address; and
request, based on the cache miss, the data from the second cache or the memory, wherein the second cache or the memory is selected for the request based on a current bandwidth utilization of the processor interconnect.