17823462. METHOD OF SUBMITTING WORK TO FABRIC ATTACHED MEMORY simplified abstract (Micron Technology, Inc.)

From WikiPatents
Jump to navigation Jump to search

METHOD OF SUBMITTING WORK TO FABRIC ATTACHED MEMORY

Organization Name

Micron Technology, Inc.

Inventor(s)

Tony M. Brewer of Plano TX (US)

METHOD OF SUBMITTING WORK TO FABRIC ATTACHED MEMORY - A simplified explanation of the abstract

This abstract first appeared for US patent application 17823462 titled 'METHOD OF SUBMITTING WORK TO FABRIC ATTACHED MEMORY

Simplified Explanation

The method described in the patent application involves receiving work packets from a separate computing device via a fabric interconnect at a command manager (CM) of a memory controller of a fabric attached memory (FAM) device. The work packets contain memory access tasks to be performed by a FAM computing resource local to the FAM device. The method includes determining the work class of the work packet, placing the work packet in a CM work queue local to the CM for the work class when space is available, and placing the work packet in a destination work queue in the memory array of the FAM device when the CM work queue is full.

  • Receiving work packets from a separate computing device via a fabric interconnect at a command manager of a memory controller of a FAM device
  • Work packets contain memory access tasks for a FAM computing resource local to the FAM device
  • Determining the work class of the work packet
  • Placing the work packet in a CM work queue when space is available
  • Placing the work packet in a destination work queue in the memory array of the FAM device when the CM work queue is full

Potential Applications

  • Distributed computing systems
  • Memory controllers
  • Fabric attached memory devices

Problems Solved

  • Efficient management of memory access tasks in distributed computing systems
  • Optimizing workload distribution in memory controllers

Benefits

  • Improved performance in distributed computing systems
  • Enhanced memory access task handling
  • Better utilization of computing resources


Original Abstract Submitted

A method performed by a distributed computing system includes receiving a work packet from a separate computing device via a fabric interconnect at a command manager (CM) of a memory controller of a fabric attached memory (FAM) device, wherein the work packet includes a memory access to be performed by a FAM computing resource local to the FAM device; determining a work class of the work packet; placing the work packet in a CM work queue local to the CM for the work class when space is available in the CM work queue for the work class; and when the CM work queue for the work class is full, placing the work packet in a destination work queue according to an address included in the work packet, wherein the destination queue is implemented in a memory array of the FAM device external to the memory controller.