Cross-Channel ECC Merging for Memory RAS Without Lockstepping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional mechanisms for achieving high memory reliability, availability, and serviceability (RAS) in computer systems are resource-inefficient, leading to performance inefficiencies and high power consumption due to requirements like lockstepping and capacity costs.

Innovation Solution

A mechanism that merges error-correction code (ECC) techniques, such as x4 Double Device Data Correction (DDDC) and x8 Single Device Data Correction (SDDC), without employing lockstepping or incurring capacity costs, by using failure handling and correction logic to detect and correct memory device failures within the ECC merging mechanism.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If rank sparing is used to achieve high memory reliability, then reliability is improved, but capacity cost increases due to requiring an entire rank of memory held in reserve

Engineering Contradiction:
Improvememory reliabilityVSAvoidcapacity cost
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent merges error correction capabilities across multiple memory channels by combining ECC locators from different channels. When a device failure is detected in one channel, the system merges the ECC locator from the failed channel with the ECC locator from another channel, allowing the combined ECC information to correct errors without requiring dedicated spare capacity in a single channel. This distributes the reliability function across channels rather than concentrating it in a reserved rank.

Inventive Principle:
Principle #5Merging (Combining)

2Reliability

If double device data correction (DDDC) with lockstepping is used to achieve high memory reliability, then reliability is improved, but power consumption increases significantly

Engineering Contradiction:
Improvememory reliabilityVSAvoidpower consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent implements a dynamic error correction approach where the system adapts its correction strategy based on the actual error patterns detected. Rather than continuously operating in a fixed lockstepping mode that consumes high power, the system dynamically selects between different correction methods (single channel correction, merged channel correction, or cross-channel merging) depending on the failure mode detected, thereby reducing power consumption while maintaining reliability.

Inventive Principle:
Principle #15Dynamics

3Reliability

If lockstepping is used to achieve high memory reliability, then reliability is improved, but bandwidth and performance are reduced

Engineering Contradiction:
Improvememory reliabilityVSAvoidbandwidth
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments the error correction function into independent channel-level operations rather than requiring synchronized lockstepping across entire memory ranks. Each channel can be corrected independently or merged with another channel as needed, allowing parallel processing of multiple channels and maintaining higher bandwidth utilization while achieving the same reliability goals through distributed correction rather than centralized lockstepping.

Inventive Principle:
Principle #1Segmentation

4Reliability

If conventional ECC schemes are used to achieve high memory reliability, then reliability is improved, but serviceability and availability are reduced due to performance penalties

Engineering Contradiction:
Improvememory reliabilityVSAvoidserviceability
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent creates a universal error correction mechanism that can handle multiple failure modes through a single merged ECC framework. The same merged ECC locator structure serves multiple functions: it can correct single-device errors within a channel, double-device errors across channels, and adapt to different failure patterns without requiring separate correction schemes. This multi-functionality improves serviceability by providing a unified approach to handling various memory failures.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS9229828B2Mechanism for achieving high memory reliability, availability and serviceability
Publication Date: 2016.01.05 INTEL CORP

AI summary

A mechanism is described for achieving high memory reliability, availability, and serviceability (RAS) according to one embodiment of the invention. A method of embodiments of the invention includes detecting a permanent failure of a first memory device of a plurality of memory devices of a first channel of a memory system at a computing system, and eliminating the first failure by merging a first error-correction code (ECC) locator device of the first channel with a second ECC locator device of a second channel, wherein merging is performed at the second channel.