Methods and apparatus for reducing accelerator-memory access costs in platforms with multiple memory channels. The apparatus includes a computing platform having multiple accelerators and multiple memory devices accessed via a plurality of memory channels. Jobs are submitted via software running on the computing platform to access a function to be offloaded to an accelerator. Under the offloaded function, the accelerator accesses one or more buffers that collectively requiring access via multiple memory channels among the plurality of memory channels. Accelerators having an available instance of the function are identified, and an aggregate cost for accessing the one or more buffers via the multiple memory channels are calculated for each of the accelerators. The accelerator with the least aggregate cost is then selected to offload the function to. New Instruction Set Architecture (ISA) instructions are also disclosed to identify memory pages and memory channels used for buffers.

A apparatus includes a storing region and an access controller. The storing region includes: a first area to which a value indicating one of first and second bit modes is set; a second area to which information indicating whether a third bit mode is enabled or disabled is set, the first and second bit modes specifying a bit length of addresses in a memory to first and second bit lengths, respectively, the third bit mode specifying the bit length to a third bit length greater than second bit length; and an address area to which an address in the first or second bit length is set. When the information in the second area indicates that the third bit mode is enabled, the access controller accesses the memory based on a concatenated address defined by concatenating the value set in the first area and the address set in the address area.

A method for accessing an extended memory, a device, and a system are disclosed. According to the method, after receiving a first memory access requests from a processor system in a computer, an extended memory controller sends a read request for obtaining to-be-accessed data to the extended memory and return, to the processor system, a first response message indicating the to-be-accessed data has not been obtained. The extended memory controller writes the to-be-accessed data into a data buffer after receiving the to-be-accessed data returned by the extended memory. After receiving, from the processor system, a second memory access request comprising a second access address, the extended memory controller returns, to the processor system, the to-be-accessed data in the data buffer in response to the second memory access request, wherein the second access address is different from the first access address and points to the physical address of the to-be-accessed data.

Techniques for setting a 2-level auto-close timer to access a memory device include examples of setting first and second time values for the 2-level auto-close timer to cause accessed rows to auto-close following a cache line access to a row of a bank of memory devices. For these examples, the cache line access is responsive to a multi-channel address interleaving policy that causes either successive or non-successive cache line accesses to the bank of memory devices.

One of a plurality of chip select inputs of a load-reduced dual inline memory module (LRDIMM) may be repurposed to an address input. One of a plurality of memory ranks of the LRDIMM may be selected based on a remainder of the plurality of chip select inputs. The repurposed chip select input may be used to support non-binary rank multiplication of the LRDIMM.

A storage system with a controller having a persistent memory interface to local memory is provided. The persistent memory can be used to store a logical-to-physical address table. A logical-to-physical address table manager, local to the controller or remote in a secondary controller, can be used to access the logical-to-physical address table. The manager can be configured to improve bandwidth and performance in the storage system.

Various embodiments include methods for implementing flexible ranks in a memory system. Embodiments may include receiving, at a memory controller, a first memory access command and a first address at which to implement the first memory access command in a logical rank, generating, by the memory controller, a first signal configured to indicate to a first memory device of the logical rank to implement the first memory access command via a first partial channel, sending, from the memory controller, the first signal to the first memory device, generating, by the memory controller, a second signal configured to indicate to a second memory device of the logical rank that is different from the first memory device to implement the first memory access command via a second partial channel, and sending, from the memory controller, the second signal to the second memory device.

In one embodiment, a host controller includes: a first input/output (I/O) buffer to couple to a first communication line of an interconnect; a second I/O buffer to couple to a second communication line of the interconnect; and a device group selection circuit to dynamically cause the first communication line to communicate a clock signal to a first device group including one or more first devices to couple to the interconnect and dynamically cause the second communication line to communicate a data signal to the first device group when a communication is to be addressed to at least one of the one or more first devices of the first device group, such that the communication is transparent to at least another device group to couple to the interconnect. Other embodiments are described and claimed.

Methods, systems, and devices for array access with receiver masking are described. A first device may issue to a second device a first sequence of write commands for a set of data. The first sequence of write commands may indicate different memory addresses in an order. After issuing the first sequence of write commands, the first device may issue to the second device a second sequence of read commands for the set of data. The second sequence of read commands may indicate the different memory addresses in the same order as the first sequence of write commands. Based on issuing the second sequence of read commands, the first device may receive the set of data from the second device.

Various embodiments include methods for implementing flexible ranks in a memory system. Embodiments may include receiving, at a memory controller, a first memory access command and a first address at which to implement the first memory access command in a logical rank, generating, by the memory controller, a first signal configured to indicate to a first memory device of the logical rank to implement the first memory access command via a first partial channel, sending, from the memory controller, the first signal to the first memory device, generating, by the memory controller, a second signal configured to indicate to a second memory device of the logical rank that is different from the first memory device to implement the first memory access command via a second partial channel, and sending, from the memory controller, the second signal to the second memory device.