Pre-Boot Surveillance Using Parallel CPU and GPU Diagnostics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current pre-boot surveillance in information handling systems lacks comprehensive monitoring, fails to utilize all processors, lacks connectivity in early phases, requires OS boot for telemetry, and prolongs diagnostic times, limiting auto-healing capabilities.
Innovation Solution
Implement a system that utilizes both a host CPU and a non-CPU processing unit for parallel execution of boot and diagnostic tasks during the pre-boot environment, enabling early initialization of Bluetooth/Wi-Fi stacks for live telemetry and cloud connectivity, and distributing diagnostic workloads across multiple processors for concurrent diagnostics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pre-boot surveillance is implemented using only the host CPU, then boot tasks can be performed, but diagnostic tasks cannot be performed in parallel and boot time increases
Solution Approach 1:
The system divides processing tasks between the host CPU (boot tasks) and non-CPU processing units (diagnostic tasks). The surveillance functionality is segmented into separate processing units that can operate independently during the pre-boot phase, enabling parallel execution without interfering with the main boot process.
Solution Approach 2:
The patent combines boot task execution and diagnostic task execution into a unified pre-boot environment that runs simultaneously on different processing units. This merging allows the system to perform both boot operations and surveillance operations in the same time window, eliminating the need to wait for OS boot to perform diagnostics.
2Reliability
If all processors are used for boot tasks, then boot completion is ensured, but no processors are available for learning boot paths and remediation
Solution Approach 1:
The system segments processor roles by designating the host CPU for boot task execution while assigning non-CPU processing units to diagnostic and learning tasks. This segmentation ensures that boot completion is maintained while simultaneously enabling surveillance and learning functions during the pre-boot phase.
Solution Approach 2:
The non-CPU processing units act as intermediaries that handle diagnostic and learning functions without interfering with the host CPU's boot task execution. These intermediary processors enable the system to learn boot paths and prepare remediation strategies while the main boot process continues uninterrupted.
3Ease of operation
If connectivity is initialized in the late DXE phase, then network stack is ready, but telemetry cannot be sent in early PEI and DXE phases
Solution Approach 1:
The system performs preliminary initialization of the network stack during the PEI phase, before the late DXE phase. This preliminary action enables the connectivity subsystem to be ready earlier, allowing telemetry data to be transmitted during early boot phases rather than waiting for the later DXE phase completion.
Data Source
AI summary
An information handling system may include a host central processing unit (CPU); and a non-CPU processing unit. The information handling system may be configured to, during execution of a pre-boot environment: perform boot tasks on the host CPU; and perform diagnostic tasks on the non-CPU processing unit. The non-CPU processing unit may be an application processing unit (APU) and/or a graphics processing unit (GPU).


