# US 2020/0394145 A1 · Memory module threading with staggered data transfers

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0394145 A1
- Document ID: US20200394145A1
- Status: Published application
- Kind: A1
- Application number: 16/914,221
- Filed: 2020-06-26
- Published: 2020-12-17
- Earliest claimed: 2012-08-10
- Assignees: Rambus Inc.
- Applicants: Rambus Inc.
- Inventors: Hongzhong Zheng; Frederick A Ware
- Web page: https://patentgrep.com/patent/US20200394145A1

## CPC

- G06F 13/1673: {using buffers}
- G06F 9/4881: {Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues}
- G06F 13/1678: {using bus width}
- G06F 13/4059: {where the synchronisation uses buffers, e.g. for speed matching between buses}
- G11C 11/4076: Timing circuits (for regeneration management G11C11/406)
- G11C 11/4094: Bit-line management or control circuits
- G06F 2209/486: Scheduler internals
- G06F 2209/5018: Thread allocation
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G11C: STATIC STORES (semiconductor memory devices H10B)

## Abstract

A method of transferring data between a memory controller and at least one memory module via a primary data bus having a primary data bus width is disclosed. The method includes accessing a first one of a memory device group via a corresponding data bus path in response to a threaded memory request from the memory controller. The accessing results in data groups collectively forming a first data thread transferred across a corresponding secondary data bus path. Transfer of the first data thread across the primary data bus width is carried out over a first time interval, while using less than the primary data transfer continuous throughput during that first time interval. During the first time interval, at least one data group from a second data thread is transferred on the primary data bus.

## Claims

### Claim 1

(canceled)

### Claim 2

A buffer circuit comprising:
a command interface to receive memory access commands from a memory controller;
a primary data bus interface having a primary data width for coupling to the memory controller, the primary data bus for transferring data at a primary data bandwidth between the buffer circuit and the memory controller;
a secondary data bus interface for coupling to a plurality of secondary buses, the secondary buses for transferring data between the buffer circuit and a plurality of memory devices, each of the secondary buses corresponding to respective threads to transfer data at a secondary data bandwidth that is less than the primary data bandwidth; and
steering logic to transfer a first data thread across the primary data width over a first time interval, while using less than the primary data bandwidth during that first time interval, and during the first time interval, transferring at least one data group from a second data thread on the primary data bus.

### Claim 3 (depends on 2)

The buffer circuit according to claim 2, wherein:
the steering logic includes circuitry to distribute portions of read data of the first data thread from a first secondary data bus to plural primary data bus paths.

### Claim 4 (depends on 2)

The buffer circuit according to claim 2, wherein:
the steering logic includes circuitry to distribute portions of write data of the first data thread from plural primary data bus paths to a secondary data bus path.

### Claim 5 (depends on 2)

The buffer circuit according to claim 2, wherein:
the command interface receives the threaded memory access commands in accordance with a dynamic random access memory (DRAM) protocol.

### Claim 6 (depends on 2)

The buffer circuit according to claim 2, wherein:
the steering logic includes circuitry to distribute a first data group of the first data thread from a first secondary data bus and a second data group from a second secondary data bus to the primary data bus such that the data groups from each data thread are staggered spatially.

### Claim 7 (depends on 6)

The buffer circuit according to claim 6, wherein:
the circuitry to distribute the first data group is configured to distribute the first data group of the first data thread across a first portion of the primary data bus width during a first time interval, and the second data group of the second data thread across a second portion of the primary data bus width during the first time interval.

### Claim 8 (depends on 2)

The buffer circuit according to claim 2, wherein:
the steering logic includes circuitry to distribute a first data group of the first data thread from a first secondary data bus and a second data group from a second secondary data bus to the primary data bus such that the data groups from each data thread are staggered temporally.

### Claim 9 (depends on 8)

The buffer circuit according to claim 8, wherein:
the circuitry to distribute the first data group is configured to distribute the first data group of the first data thread across the entire primary data bus width during a first time interval, and configured to interleave the first data group with the second data group across the entire primary data bus width during a second time interval.

### Claim 10

A method of operation in a buffer circuit, comprising:
receiving memory access commands from a memory controller via a command interface;
transferring data at a primary data bandwidth between the buffer circuit and the memory controller via a primary data bus interface having a primary data width;
transferring data between the buffer circuit and a plurality of memory devices via a secondary data bus interface, the secondary bus interface for coupling to a plurality of secondary buses, each of the secondary buses corresponding to respective threads to transfer data at a secondary data bandwidth that is less than the primary data bandwidth; and
steering transfer of a first data thread across the primary data width over a first time interval with steering logic, while using less than the primary data bandwidth during that first time interval, and during the first time interval, steering transfer of at least one data group from a second data thread on the primary data bus.

### Claim 11 (depends on 10)

The method according to claim 10, wherein the steering transfer of the first data thread comprises:
distributing portions of read data of the first data thread from a first secondary data bus to plural primary data bus paths.

### Claim 12 (depends on 10)

The method according to claim 10, wherein the steering transfer of the first data thread comprises:
distributing portions of write data of the first data thread from plural primary data bus paths to a secondary data bus path.

### Claim 13 (depends on 10)

The memory according to claim 10, wherein the steering transfer of the first data thread comprises:
distributing a first data group of the first data thread from a first secondary data bus and a second data group from a second secondary data bus to the primary data bus such that the data groups from each data thread are staggered spatially.

### Claim 14 (depends on 13)

The method according to claim 13, wherein the distributing comprises:
distributing the first data group of the first data thread across a first portion of the primary data bus width during a first time interval, and the second data group of the second data thread across a second portion of the primary data bus width during the first time interval.

### Claim 15 (depends on 10)

The method according to claim 10, wherein the steering transfer of the first data thread comprises:
distributing a first data group of the first data thread from a first secondary data bus and a second data group from a second secondary data bus to the primary data bus such that the data groups from each data thread are staggered temporally.

### Claim 16 (depends on 15)

The method according to claim 15, wherein the distributing comprises:
distributing the first data group of the first data thread across the entire primary data bus width during a first time interval, and interleaving the first data group with the second data group that is distributed across the entire primary data bus width during a second time interval.

### Claim 17

A buffer integrated circuit (IC) chip, comprising:
a primary data bus interface having a primary data width for coupling to a memory controller, the primary data bus for transferring data at a primary data bandwidth between the buffer circuit and the memory controller;
a secondary data bus interface for coupling to a plurality of secondary buses, the secondary buses for transferring data between the buffer circuit and a plurality of memory devices, each of the secondary buses corresponding to respective threads to transfer data at a secondary data bandwidth that is less than the primary data bandwidth; and
steering logic to transfer a first data thread across the primary data width over a first time interval, while using less than the primary data bandwidth during that first time interval, and during the first time interval, transferring at least one data group from a second data thread on the primary data bus.

### Claim 18 (depends on 17)

The buffer IC chip according to claim 17, wherein the steering logic comprises:
circuitry to spatially stagger a first data group of the first data thread with the at least one data group of the second data thread.

### Claim 19 (depends on 18)

The buffer IC chip according to claim 18, wherein the circuitry to spatially stagger further comprises:
circuitry to distribute the first data group of the first data thread across a first portion of the primary data bus width during a first time interval, and the second data group of the second data thread across a second portion of the primary data bus width during the first time interval.

### Claim 20 (depends on 17)

The buffer IC chip according to claim 17, wherein the steering logic comprises:
circuitry to temporally stagger a first data group of the first data thread with the at least one data group of the second data thread.

### Claim 21 (depends on 20)

The method according to claim 20, wherein the circuitry to temporally stagger further comprises:
circuitry to distribute the first data group of the first data thread across the entire primary data bus width during a first time interval, and to interleave the first data group with the at least one data group of the second data thread across the entire primary data bus width during a second time interval.
