skills.rest
JORDAN0615
JORDAN0615COMMUNITY
Explore all skills by this author →
GitHub

jetson-optimize-memory

Reclaim unused DRAM by disabling peripheral carveouts across Jetson boot layers.

1|Updated Jun 12, 2026
Search Tags:#bsp#embedded linux#jetson#dram optimization#carveout#swiotlb#memory reclaim
One-click install
npx skills add https://github.com/JORDAN0615/bsp_factory --skill jetson-optimize-memory
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: jetson-optimize-memory
Source: https://github.com/JORDAN0615/bsp_factory/tree/main/skills/jetson-optimize-memory
Command: npx skills add https://github.com/JORDAN0615/bsp_factory --skill jetson-optimize-memory

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Jetson BSPs reserve DRAM for display, camera, and DMA subsystems even when those peripherals are not used in headless or no-camera deployments, wasting valuable memory that could be allocated to application workloads.

Core Features & Use Cases

  • Cross-layer DRAM reclamation: Disables unused DRAM carveouts across all four boot layers (MB1 BCT, MB2 BCT, kernel reserved-memory, SWIOTLB) to maximize freed memory.
  • Validated scenario recipes: Provides pre-tested, chip-specific configuration sets for T234 (Orin) and T264 (Thor) platforms for headless, no-camera, and SWIOTLB optimization use cases.
  • Use case example: A headless edge AI inference deployment running on a Jetson Orin Nano can reclaim hundreds of megabytes of DRAM by disabling unused display and camera subsystems, freeing memory for larger AI models or higher throughput workloads.

Quick Start

Use the jetson-optimize-memory skill with the 'headless' argument to disable unused display and camera DRAM carveouts on your Jetson Orin deployment to free up memory for your workload.

Frequently Asked Questions about jetson-optimize-memory

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reclaim unused DRAM on Jetson Orin for headless deployments?▼

Reclaim unused DRAM on Jetson Orin by disabling firmware and kernel-level carveouts across MB1 BCT, MB2 BCT, kernel reserved-memory, and SWIOTLB boot layers to free memory for application workloads.

Why does my Jetson BSP reserve memory for cameras when no camera is connected?▼

Jetson BSPs reserve DRAM for display, camera, and DMA subsystems by default, even when peripherals are unused in headless or no-camera deployments, wasting valuable memory that could be allocated to workloads.

Can I disable SWIOTLB carveout on T234 Jetson platforms?▼

Yes, you can disable SWIOTLB carveout on T234 Jetson platforms using validated, chip-specific configuration sets designed for SWIOTLB optimization scenarios to safely reduce DRAM waste.

What is the best way to free up memory for edge AI models on Jetson Thor?▼

The best way to free up memory for edge AI models on Jetson Thor is cross-layer DRAM reclamation, which disables unused display and camera carveouts across all four boot layers using pre-tested recipes.

How to prevent boot failures when modifying Jetson kernel reserved-memory?▼

Prevent boot failures when modifying Jetson kernel reserved-memory by adhering strictly to validated recipe sets and performing mandatory pre-flash verification and post-boot validation checks.

Related Skills

Explore More Software Engineering Skills→
NVIDIA-AI-IOTNVIDIA-AI-IOT
46

jetson-optimize-memory

Disable unused subsystems to reclaim DRAM in Jetson deployments.

Official
Advanced
NVIDIA-AI-IOTNVIDIA-AI-IOT
110

jetson-memory-audit

Measure Jetson DRAM and NvMap usage on live NVIDIA Jetson devices.

Official
Advanced
FastLEDFastLED
7.5k

driver-review

Review embedded driver changes for DMA, interrupt, and peripheral access safety.

Official
Advanced
JORDAN0615JORDAN0615

jetson-quick-start

Guide Jetson BSP workspace setup via a routing-critical questionnaire.

Community
Advanced