Composable infrastructure enabled by heterogeneous architecture, delivered by CXL based cached switch SOC
Described herein are systems, methods, and products utilizing a cache coherent switch on chip. The cache coherent switch on chip may utilize Compute Express Link (CXL) interconnect open standard and allow for multi-host access and the sharing of resources. The cache coherent switch on chip provides for resource sharing between components while independent of a system processor, removing the system processor as a bottleneck. Cache coherent switch on chip may further allow for cache coherency between various different components. Thus, for example, memories, accelerators, and/or other components within the disclose systems may each maintain caches, and the systems and techniques described herein allow for cache coherency between the different components of the system with minimal latency.
1. A system comprising:
a first server device comprising:
a first memory device; and
a first cache coherent switch on chip, communicatively coupled to the first memory device via a Compute Express Link (CXL) protocol; and
a second server device, communicatively coupled to the first server device via a data connection, the second server device comprising:
a second memory device; and
a second cache coherent switch on chip, communicatively coupled to the second memory device via the CXL protocol and communicatively coupled to the first cache coherent switch on chip by the data connection via the CXL protocol, wherein second cache coherent switch on chip is configured to:
determine that the second memory device is to be shared with the first server device;
generate an access key for the first server device to access the second memory device; and
provide, to the first server device via the data connection, the access key; and
wherein the first cache coherent switch is configured to communicate CXL protocol memory commands to the second cache coherent switch, each of the CXL protocol memory commands including an instruction and the access key for the first server device to access the second memory device.
2. The system of claim 1 , wherein the first cache coherent switch on chip is configured to:
receive the access key; and
provide, to the second cache coherent switch on chip, read or write instructions for the second memory device, wherein the read or write instructions comprise the access key and a request for reading or writing first data.
3. The system of claim 2 , wherein the second cache coherent switch on chip is further configured to:
receive the read or write instructions;
read or write, in response to receiving the read or write instructions, the first data; and
provide, to the first cache coherent switch on chip and based on the reading or writing the first data, a read or write acknowledgement.
4. The system of claim 3 , wherein the read or write acknowledgement is configured to maintain cache coherency within the system.
5. The system of claim 1 , wherein the first cache coherent switch on chip and the second cache coherent switch on chip are further configured to share cached memory within the first memory device and the second memory device.
6. The system of claim 5 , wherein the first cache coherent switch on chip is configured to:
receive a request for recall of cached first data;
determine that the cached first data is not available within the first memory device; and
provide, to the second cache coherent switch on chip, a request for the cached first data.
7. The system of claim 6 , wherein the second cache coherent switch on chip is further configured to:
determine that the cached first data is stored within the second memory device; and
provide, to the first cache coherent switch on chip, the cached first data.
8. The system of claim 6 , wherein the second server device further comprises:
a microprocessor; and
a third memory device communicatively coupled to the microprocessor, wherein the second cache coherent switch on chip is further configured to:
determine that the cached first data is not stored within the second memory device; and
provide, to the microprocessor, the request for the cached first data from the third memory device;
receive the cached first data from the microprocessor; and
provide, to the first cache coherent switch on chip, the cached first data.
9. The system of claim 1 , further comprising:
a third server device, communicatively coupled to the first server device and the second server device via the data connection, the third server device comprising:
a third memory device; and
a third cache coherent switch on chip, communicatively coupled to the third memory device via the CXL protocol and communicatively coupled to the first cache coherent switch on chip and the second cache coherent switch on chip by the data connection via the CXL protocol, wherein the first memory device, the second memory device, and the third memory device are shared between the first cache coherent switch on chip, the second cache coherent switch on chip, and the third cache coherent switch on chip.
10. The system of claim 9 , wherein the first cache coherent switch on chip is configured to:
determine that a requested cached first data is not stored within the first memory device; and
provide, to the second cache coherent switch on chip and the third cache coherent switch on chip, a request for the cached first data.
11. The system of claim 9 , wherein the first cache coherent switch on chip is configured to:
receive, based on the request for the cached first data, erasure data; and
construct the cached first data by replacing missing data blocks with the erasure data.
12. The system of claim 1 , wherein the first memory device and the second memory device are DRAM or persistent memory.