A computer-implemented method includes receiving, by a processing unit, from a first tenant, a query to retrieve data from a nonrelational database system. The method further includes determining, by the processing unit, that an index associated with the query is cached in a shared index cache, wherein the shared index cache stores indexes for a plurality of tenants. The method further includes retrieving, by the processing unit, a result of the query based on the index in the shared index cache. The method further includes outputting, by the processing unit, the result of the query.

Methods, systems, and devices for cache architectures for memory devices are described. For example, a memory device may include a main array having a first set of memory cells, a cache having a second set of memory cells, and a cache delay register configured to store an indication of cache addresses associated with recently performed access operations. In some examples, the cache delay register may be operated as a first-in-first-out (FIFO) register of cache addresses, where a cache address associated with a performed access operation may be added to the beginning of the FIFO register, and a cache address at the end of the FIFO register may be purged. Information associated with access operations on the main array may be maintained in the cache, and accessed directly (e.g., without another accessing of the main array), at least as long as the cache address is present in the cache delay register.

Methods, systems, and devices for cache architectures for memory devices are described. For example, a memory device may include a main array having a first set of memory cells, a cache having a second set of memory cells, and a cache delay register configured to store an indication of cache addresses associated with recently performed access operations. In some examples, the cache delay register may be operated as a first-in-first-out (FIFO) register of cache addresses, where a cache address associated with a performed access operation may be added to the beginning of the FIFO register, and a cache address at the end of the FIFO register may be purged. Information associated with access operations on the main array may be maintained in the cache, and accessed directly (e.g., without another accessing of the main array), at least as long as the cache address is present in the cache delay register.

An integrated circuit includes logic circuits, first buffer circuits coupled to external ports of the integrated circuit, second buffer circuits that are each coupled to one of the logic circuits, and a crossbar circuit coupled to the first and the second buffer circuits. The crossbar circuit is configurable to provide data transfer between the logic circuits and the external ports of the integrated circuit through the first buffer circuits and the second buffer circuits.

An integrated circuit includes logic circuits, first buffer circuits coupled to external ports of the integrated circuit, second buffer circuits that are each coupled to one of the logic circuits, and a crossbar circuit coupled to the first and the second buffer circuits. The crossbar circuit is configurable to provide data transfer between the logic circuits and the external ports of the integrated circuit through the first buffer circuits and the second buffer circuits.

A method of operating a storage device, including a first memory region having a lowest bit density, a second memory region having a medium bit density, and a third memory region having a highest bit density, includes determining a hotness of a logical address received with a write command and data to be written, from a host, based on the determined hotness being greater than a first hotness threshold, determining whether a wear level of the first memory region is greater than a wear threshold, and increasing the first hotness threshold and storing the data in the second memory region based on the wear level of the first memory region being greater than a threshold.

A method of operating a storage device, including a first memory region having a lowest bit density, a second memory region having a medium bit density, and a third memory region having a highest bit density, includes determining a hotness of a logical address received with a write command and data to be written, from a host, based on the determined hotness being greater than a first hotness threshold, determining whether a wear level of the first memory region is greater than a wear threshold, and increasing the first hotness threshold and storing the data in the second memory region based on the wear level of the first memory region being greater than a threshold.

A method and apparatus physically partitions clean and dirty cache lines into separate memory partitions, such as one or more banks, so that during low power operation, a cache memory controller reduces power consumption of the cache memory containing the clean only data. The cache memory controller controls refresh operation so that data refresh does not occur for clean data only banks or the refresh rate is reduced for clean data only banks. Partitions that store dirty data can also store clean data, however other partitions are designated for storing only clean data so that the partitions can have their refresh rate reduced or refresh stopped for periods of time. When multiple DRAM dies or packages are employed, the partition can occur on a die or package level as opposed to a bank level within a die.

This application describes a hardware accelerator, a computer system, and a method for accelerating Graph Neural Network (GNN) node attribute fetching. The hardware accelerator comprises a GNN attribute processor; and a first memory, wherein the GNN attribute processor is configured to: receive a graph node identifier; determine a target memory address within the first memory based on the graph node identifier; determine, based on the received graph node identifier, whether attribute data corresponding to the received graph node identifier is cached in the first memory at the target memory address; and in response to determining that the attribute data is not cached in the first memory: fetch the attribute data from a second memory, and write the fetched attribute data into the first memory at the target memory address.

This application describes a hardware accelerator, a computer system, and a method for accelerating Graph Neural Network (GNN) node attribute fetching. The hardware accelerator comprises a GNN attribute processor; and a first memory, wherein the GNN attribute processor is configured to: receive a graph node identifier; determine a target memory address within the first memory based on the graph node identifier; determine, based on the received graph node identifier, whether attribute data corresponding to the received graph node identifier is cached in the first memory at the target memory address; and in response to determining that the attribute data is not cached in the first memory: fetch the attribute data from a second memory, and write the fetched attribute data into the first memory at the target memory address.