Disaggregated Crash Log Collection With Distributed Dielet Storage

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Legacy crash log solutions are limited to single integrated circuit dies and fail to scale for multi-die next generation products, losing crash log records during global resets and lacking accessibility across disaggregated dielets.

Innovation Solution

A disaggregated crash log collection architecture that includes distributed crash log collection across dielets with dedicated storage and communication links, utilizing a central agent for coordinated data retrieval post-reboot.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If legacy crash log architecture is used on a single integrated circuit die, then crash log collection is simple and straightforward, but it cannot scale to multi-die next generation products and loses crash log records during global resets

Engineering Contradiction:

Improvescalability to multi-die productsVSAvoiddata retention during global reset

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The crash log collection architecture is segmented across multiple dielets, with each dielet having its own local crash log storage. This segmentation allows the system to scale from single-die to multi-die configurations while maintaining crash log data independently on each dielet, preventing data loss during global resets.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

A package-level crash log collection mechanism acts as an intermediary between individual dielets and the external world. This intermediary collects crash logs from multiple dielets through interconnect links and provides a unified interface, enabling scalable multi-die support while preserving data integrity across reset events.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Adaptability or versatility

If crash log storage is distributed across multiple dielets, then scalability and data retention are improved, but system complexity increases

Engineering Contradiction:

Improvescalability to multi-die productsVSAvoiddistributed storage and communication architecture

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The crash log storage and communication infrastructure is designed with multi-functionality to handle both single-die and multi-die scenarios. The same basic building blocks (crash log storage, detection logic, communication interfaces) serve multiple purposes across different dielet configurations, reducing overall system complexity despite the distributed architecture.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The crash log collection architecture uses a nested structure where individual dielet-level crash log mechanisms are embedded within a package-level collection system. Each dielet contains its own crash log storage and detection logic, which are then aggregated by the package-level mechanism, creating a scalable nested architecture that manages complexity through hierarchical organization.

Inventive Principle:

Principle #7Nested doll (Nesting)

3Reliability

If dedicated crash log storage is implemented on each dielet, then crash log data is retained during resets, but device complexity and manufacturing cost increase

Engineering Contradiction:

Improvecrash log data retentionVSAvoidmanufacturing of multi-die product

Core Design Contradiction:

ReliabilityVSEase of manufacture

Solution Approach 1:

Each dielet is designed with homogeneous crash log storage capabilities using identical or similar storage structures and detection logic. This homogeneity simplifies manufacturing by allowing standardization of crash log components across all dielets, reducing variability and easing the manufacturing process despite the multi-die complexity.

Inventive Principle:

Principle #33Homogeneity

Data Source

PatentUS20250272174A1Crash log collection architecture for disaggregated die products

Publication Date: 2025.08.28 INTEL CORP

AI summary

Methods and apparatus relating to a crash log collection architecture for disaggregated die products are described. In one embodiment, an integrated circuit device (e.g., a semiconductor package such as an SOC, NOP, SOP, and so on) includes a plurality of dielets coupled via one or more communication links. Two or more dielets from the plurality of dielets each include a storage device to store a crash log. Logic circuitry on at least one dielet from the plurality of dielets causes storage of the crash log in the storage device in response to detection of a failure. Other embodiments are also disclosed and claimed.