Multi-Die ECC Architecture for High RBER Memory Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Error Correction Code (ECC) schemes in non-volatile memory systems lack sufficient resilience to Raw Bit Error Rate (RBER), as they rely on XOR operations for data recovery, which is inadequate when multiple memory dies fail, leading to reduced data integrity.
Innovation Solution
Implementing a super-BCH codeword scheme where the 20th die stores the parity of the super-BCH codeword instead of the XOR result, allowing for enhanced error correction capabilities by using a second error correction scheme to encode the parity portion, thereby improving resilience to RBER by a factor of 10 compared to traditional XOR-based methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If XOR-based data recovery is used in traditional ECC schemes, then data recovery is simple and fast, but resilience to RBER is insufficient when multiple memory dies fail
Solution Approach 1:
The patent segments the error correction capability by introducing a hierarchical structure: traditional ECC is applied to each individual memory die, while a super-BCH code is applied across multiple dies. This segmentation allows each layer to handle errors independently, with the super-BCH layer providing additional protection when multiple dies fail, thereby improving RBER resilience without requiring complete redesign of the entire error correction system.
Solution Approach 2:
The patent implements nested error correction by embedding traditional ECC codes within a broader super-BCH code structure. The super-BCH code encompasses multiple memory dies and their individual ECC-protected codewords, creating a nested hierarchy where inner ECC layers handle die-level errors and outer super-BCH layers handle multi-die failures. This nesting enables progressive error correction without increasing overall system complexity significantly.
2Reliability
If super-BCH codeword scheme is implemented with second error correction scheme, then resilience to RBER improves by factor of 10, but device complexity increases
Solution Approach 1:
The patent applies preliminary error correction by first using traditional ECC to correct errors within each individual memory die before the data is combined into super-BCH codewords. This preliminary correction reduces the error burden on the super-BCH decoding stage, allowing it to focus on correcting residual errors across multiple dies. This two-stage preliminary action improves data integrity while managing decoding complexity by pre-processing errors at the die level.
Solution Approach 2:
The patent introduces an intermediary layer of super-BCH encoding that mediates between individual die-level ECC codes and the final data recovery process. The super-BCH code acts as an intermediary structure that combines multiple ECC-protected codewords and provides an additional layer of error correction. This intermediary approach enables the system to achieve 10x improved RBER resilience while keeping the complexity manageable by distributing the correction burden across multiple hierarchical levels rather than requiring a single complex correction mechanism.
3Reliability
If traditional ECC with single error correction layer is used, then device complexity is low, but ability to correct multiple bit errors across multiple dies is insufficient
Solution Approach 1:
The patent segments error correction into two distinct functional segments: die-level ECC correction handling individual die errors, and super-BCH correction handling multi-die errors. This segmentation allows each segment to be optimized for its specific error type, with traditional ECC providing fast single-die correction and super-BCH providing comprehensive multi-die protection. The segmentation improves error correction capability while maintaining relatively low complexity by avoiding a single monolithic complex correction system.
Solution Approach 2:
The patent creates a composite error correction system by combining traditional ECC codes with super-BCH codes in a hierarchical structure. Similar to composite materials in engineering, this composite approach combines the strengths of different error correction techniques: the simplicity and speed of traditional ECC with the powerful multi-error correction capability of super-BCH codes. This composite structure enables the system to correct both single-bit and multi-bit errors across multiple dies while managing overall complexity through modular integration.
Data Source
AI summary
Described is an apparatus which comprises: a first encoder to encode data with a first error correction scheme to generate a set of codewords, each codeword of the set having a data portion and a corresponding parity portion, and each codeword of the set to be stored in a separate memory bank of a memory block; and a second encoder to encode the data portions of each codeword of the set with a second error correction scheme, the second encoder to generate a combined codeword having a data portion and a corresponding parity portion, wherein the corresponding parity portion of the combined codeword is to be stored in an additional memory bank of the memory block.


