Memory module decoder

Patent No. US7619912 (titled "Memory module decoder") on Sep 27, 2007. The application was issued on Nov 17, 2009.

What is this patent about?

’912 is related to the field of high-density memory modules, specifically those utilizing double-data-rate (DDR) dynamic random-access memory (DRAM) devices. In modern computing, memory controllers are often limited by the number of physical ranks or chip-select signals they can support, which in turn restricts the total memory capacity available per slot. The invention addresses the need to bypass these hardware limitations by allowing a computer system to interface with a high-capacity module as if it were a standard, lower-density component.

The underlying idea behind ’912 is the implementation of a virtual memory architecture where a logic element acts as an intermediary between the host computer and a larger-than-expected array of memory devices. By intercepting standard control signals—such as chip-selects and address bits—the logic element performs rank multiplication or density simulation. This allows the system to address a greater number of physical memory ranks than the memory controller is natively configured to handle, effectively tricking the BIOS into seeing a single high-density rank where multiple lower-density ranks actually exist.

The claims of ’912 focus on a memory module comprising a printed circuit board populated with a plurality of DDR memory devices arranged in a first number of ranks, managed by a circuit containing a logic element, a register, and a phase-lock loop (PLL). The independent claims specifically cover the mechanism where the logic element receives a set of input signals (chip-selects, bank addresses, and row/column addresses) corresponding to a smaller number of ranks and translates them into a larger set of output signals that control the actual, higher number of physical ranks. This translation includes responding to host commands by selectively activating specific ranks or pairs of ranks to ensure the module behaves as a unified, higher-capacity entity.

In practice, the invention works by using the logic element to re-map address and command signals in real-time. For example, an extra row address bit from the host can be converted into a column address bit or used to toggle between physical ranks that the host cannot see. To maintain signal integrity during these complex operations, the module utilizes a PLL for precise clock distribution and a register to buffer control signals. The invention also describes methods for managing data strobe (DQS) signals, such as using isolation switches or current-limiting resistors to prevent signal collisions when multiple physical devices are tied to a single logical data path.

This approach differs from prior solutions by moving the complexity of rank management off the motherboard and onto the module itself. Traditional modules are limited by the fixed number of chip-select lines provided by the memory controller; however, ’912 enables density multiplication without requiring hardware changes to the host system. By reporting a virtual configuration through the Serial Presence Detect (SPD) device, the module ensures compatibility while utilizing lower-cost, lower-density DRAM components to achieve high-capacity results that would otherwise be economically or technically unfeasible.

How does this patent fit in bigger picture?

Technical Landscape

In the mid-2000s when ’912 was filed, computer systems were typically designed with rigid memory controller architectures that supported a limited number of physical memory ranks per slot. At a time when memory capacity was typically increased by migrating to higher-density DRAM components, hardware constraints in server and personal computer motherboards often restricted the number of chip-select signals available to address these components. When systems commonly relied on a direct one-to-one mapping between the controller's control signals and the physical memory ranks, increasing memory density beyond standard architectural limits was non-trivial due to the fixed electrical and logical interface of the memory bus.

Prosecution Position

The disclosed invention represents a technical advancement in memory module architecture by integrating a logic element that performs signal translation between the host computer and a larger-than-standard array of memory devices. This architectural shift allows a memory module to present a virtualized configuration to the host, where the logic element receives a set of control signals corresponding to a smaller number of ranks and generates an expanded set of output signals to manage a higher number of physical ranks. This integration enables the use of multiple lower-density, cost-effective memory devices to emulate the behavior of fewer, higher-density devices, thereby overcoming the physical addressing constraints of the host memory controller while increasing total module capacity.

Claims

This patent contains 51 claims, with independent claims 1, 15, 28, and 39 focusing on a memory module equipped with a circuit comprising a logic element, a register, and a phase-lock loop device designed to translate input control signals from a computer system into output signals that manage a larger number of physical memory ranks than the system expects. The dependent claims further specify the hardware implementation of the logic element, the physical arrangement and spacing of memory devices on the printed circuit board, the handling of specific command types like refresh or read/write signals, and the buffering or storage of address signals during row and column access procedures to facilitate density simulation.

Key Claim Terms New

Definitions of key terms used in the patent claims.

Term (Source)Support for SpecificationInterpretation
Chip-select signal
(Claim 1, Claim 15, Claim 28, Claim 39)
Examples of such control signals include, but are not limited to, rank-select signals, also called chip-select signals. The logic element receives at least one chip-select signal from the computer system to generate a first number of chip-select signals for the physical ranks on the module.A control signal used by the computer system to select or activate specific ranks of memory devices on the memory module.
Double-data-rate (DDR) memory devices
(Claim 1, Claim 15, Claim 28, Claim 39)
The memory module comprises a plurality of double-data-rate (DDR) memory devices mounted to the printed circuit board. These devices typically have a memory configuration with a unique combination of rows, columns, and banks. The DRAM devices are generally arranged as ranks or rows of memory.A specific type of dynamic random-access memory (DRAM) that transfers data on both the rising and falling edges of the clock signal, arranged in banks and ranks on the printed circuit board.
Logic element
(Claim 1, Claim 15, Claim 28, Claim 39)
The logic element receives a set of input control signals from the computer system corresponding to a second number of memory devices. It generates a set of output control signals corresponding to a first, larger number of memory devices. This allows the memory module to utilize a higher density of physical memory devices than the computer system is configured to support.A component within a circuit or integrated circuit that receives input control signals (such as address and chip-select signals) and generates corresponding output signals to manage a larger number of memory devices or ranks than the computer system originally addresses.
Phase-lock loop device
(Claim 1, Claim 15, Claim 28, Claim 39)
The memory module includes a phase-lock loop device mounted to the printed circuit board. The phase-lock loop device is operatively coupled to the plurality of DDR memory devices, the logic element, and the register. In some embodiments, it is part of at least one integrated circuit element.A circuit component mounted on the memory module that is operatively coupled to the memory devices, logic element, and register to manage timing or clock synchronization.
Ranks
(Claim 1, Claim 15, Claim 28, Claim 39)
The DRAM devices of a memory module are generally arranged as ranks or rows of memory, each rank of memory generally having a bit width. During operation, the ranks of a memory module are selected or activated by control signals that are received from the processor. Most computer and server systems support one-rank and two-rank memory modules.Distinct rows or groupings of memory devices on a module that are selected or activated by control signals, such as chip-select signals, to be accessed by the processor.

Litigation Cases New

US Latest litigation cases involving this patent.

Case NumberFiling DateTitle
2:23-cv-00628Dec 22, 2023Netlist, Inc. v. Micron Technology, Inc. et al

Patent Family

Patent Family

File Wrapper

The dossier documents provide a comprehensive record of the patent's prosecution history - including filings, correspondence, and decisions made by patent offices - and are crucial for understanding the patent's legal journey and any challenges it may have faced during examination.

  • Get instant alerts for new documents

US7619912

Application Number
US11862931A
Filing Date
Sep 27, 2007
Publication Date
Nov 17, 2009
External Links
Slate, USPTO , Google Patents