NVIDIA-Rendering-for-Carestream-Kodak

An enterprise-grade infrastructure optimization framework designed to force, manage, and audit local and virtualized hardware acceleration across medical imaging networks. This repository contains the complete deployment scripts, self-healing runtimes, network optimizations, and analytics tracking structures to route Carestream (CS) Imaging Suite 3.26 processing pipelines directly through dedicated NVIDIA GPU hardware pipelines.

---

🛠️ Repository Architecture

This repository is organized into distinct functional layers to facilitate deployment by Hospital IT and Clinical Engineering teams:

├── .github/workflows/
│ └── psscriptanalyzer-validation.yml # CI/CD Static Code Quality Analysis Pipeline
├── dashboards/
│ └── VirusTC-Compliance-Dashboard.html # Intranet Fleet Monitor & Analytics UI
├── docs/
│ ├── nvidia-rtx6000-csimaging.md # Core Workstation HW Acceleration Mapping
│ ├── IT-NVIDIA-NVML-Workstation.md # Low-Level VRAM Allocation & Monitoring
│ ├── IT-Citrix-vGPU-Carestream.md # Datacenter Hypervisor & vGPU Provisioning
│ ├── IT-Network-DICOM-QoS.md # 9000 MTU Jumbo Frames & QoS Mapping Policies
│ ├── GIT-Version-Control-Strategy.md # Trunk-Based Change-Control Workflow
│ └── TUF-[01-05]-Endpoint-Profiles.md # Multi-Method Remote Office Viewport Guidelines
└── scripts/
├── Citrix-TUF-Endpoint-Opt.ps1 # Master PVS/MCS Image Optimization Script
├── VirusTC-SelfHealing-Auditor.ps1 # Active GPO Startup Compliance Auditor & Repair Loop
├── VirusTC-Monthly-Report-Gen.ps1 # Automated 30-Day Executive KPI Compiler
├── VirusTC-Monthly-Email-Sub.ps1 # CIO Office HTML Automated Delivery Engine
├── VirusTC-API-Listener.ps1 # Centralized REST Web API Handshake Listener
└── VirusTC-Endpoint-Scrub.ps1 # Secure Decommissioning Vault-Wiping Tool

📈 Operational Framework Overview

The architecture divides hospital endpoints into two primary functional categories, providing specialized optimizations for each:

1. Central Server / Acquisition Nodes (NVIDIA RTX Pro 6000)

Forces the offloading of raw mathematical transformations from the central CPU to dedicated GPU pipelines. This optimizes high-tier clinical processing functions:

• Ultra-Low-Dose AI Denoising: Utilizes specialized Tensor Cores to scrub quantum mottle, enabling up to a 40% reduction in patient exposure (kVp/mAs).
• 3D CBCT Iterative Reconstruction: Employs CUDA Multi-Pass Iterative loops to eliminate beam hardening and metal streak artifacts.
• Clinical Modules Acceleration: Drives Carestream Eclipse AI extensions including Automated Bone Suppression, Virtual SmartGrid math, and Enhanced Tube/Line Visualization.

2. Remote Office / Endpoints (ASUS TUF GeForce RTX 50 Series)

Optimizes client machines that lack direct physical tethers to X-ray machines. These office endpoints process heavy graphical streams pushed from the datacenter over a virtualized architecture:

• Hardware Decoding Offload: Forces the endpoint NVDEC hardware chip to decompress incoming Citrix HDX bitstreams, preserving 4:4:4 Chroma Subsampling for crisp grayscale transitions.
• Industrial-Grade Reliability: Leverages over-engineered hardware properties (military-grade 20K capacitors and dual ball bearing fans) to maintain a zero-throttling display state.

🎛️ Carestream VuePACS & DryView 5950 Calibration Matrix

To prevent visual drift between the on-screen NVIDIA hardware acceleration rendering pipeline and physical print media outputs, clinical networks must implement rule-based media sorting profiles.

1. DICOM Routing Rules & Tray Allocation

Configure the Carestream VuePACS Print Client Engine (SOP Class UID 1.2.840.10008.5.1.1.18) to separate output workflows:

• Tray 1 (Baseline): Routes standard print payloads to Blue Base DryView Laser Imaging Film (DVB).
• Tray 2 (Premium): Routes insurance/patient-paid visual upgrades to Clear Base Film cartridges for maximum back-lit density on clinical light-boards.
• Automation Hook: Sorting rules trigger dynamically when matching the custom billing macro identifier string: Premium_Clear_Sheet.

2. FDA-Compliant Adobe Look-Up Table (LUT) Calibration

To preserve Adobe-layered vector transparency profiles and match Grayscale Standard Display Function (GSDF) limits:

• Clear Base Profile: Force a rigid minimum density limit of Dmin = 0.20 to prevent background clouding under intense light-board illumination.
• Blue Base Profile: Force a rigid maximum density limit of Dmax = 3.20 to ensure deep, clean black clipping boundaries.
• Interpolation: Enforce the Linear Interpolation Curve inside the DryView Web Portal console.

---

🚦 End-to-End Infrastructure Deployment Sequence (Updated)

The updated infrastructure deployment progression must follow this strict sequence:

[1. Hypervisor / Network Setup]

➔ [2. Master PVS/MCS Image Build (Citrix-TUF-Endpoint-Opt.ps1)]

➔ [3. Standalone Hardware Provisioning (Local-TUF-Endpoint-Opt.ps1)]

➔ [4. API Listener Node Deploy]

➔ [5. Active Directory GPO & WMI Target Configuration]

---

🚦 End-to-End Infrastructure Deployment Sequence

To implement this optimization framework within a production hospital network, engineers should execute the setup steps in the following order: [1. Hypervisor / Network Setup] ➔ [2. Master PVS/MCS Image Build] ➔ [3. API Listener Node Deploy] ➔ [4. AD GPO Configuration]

Step 3: Standalone Hardware Provisioning (Non-Citrix Local Endpoints)

For physical workspace environments equipped with dedicated ASUS TUF Gaming GeForce RTX 50 Series hardware components:

1. Execute scripts/Local-TUF-Endpoint-Opt.ps1 locally or via startup automation.
2. This script enforces high-performance power configurations, sets an FDA-stabilized Windows TdrDelay (10 seconds), pins the Carestream Vue app layer exclusively onto local NVDEC hardware bitstream decoders, and executes an automated post-deployment cleanup cycle to clear scratch files and memory logs.
3. Hook the REST function to point payloads directly back to your secure VirusTC-API-Listener.ps1 pipeline over port 8443.

Step 5: Active Directory GPO & WMI Target Configuration

1. Group Policy Objects deploying local execution scripts must link the following scoping filter to prevent execution on unapproved virtual endpoints:

   SELECT * FROM Win32_VideoController WHERE Name LIKE "%NVIDIA%" OR Name LIKE "%ASUS%"

2. Import NvidiaCarestreamStabilization.admx into your Active Directory Central Store to enforce state baseline evaluations over GPO intervals.

Step 1: Hypervisor and Network Core Configuration

1. Open your datacenter core and edge switches. Enable Jumbo Frames with an MTU ceiling of 9000 bytes across all imaging VLANs.
2. In your hypervisor console (e.g., VMware vSphere), map your distributed switch MTU to 9000. Attach your physical RTX Pro 6000 hardware units to the virtual machines via a dedicated Q-Profile (Virtual Workstation) tier (NVIDIA-RTX-PRO-6000-12Q or 24Q).
3. Enforce 100% Locked Memory Reservation on the guest VMs. CUDA processing will remain disabled inside a virtual instance if memory overcommitment is permitted.

Step 2: Build the Golden Master Image

1. Boot the target Master Image VM in Private/Maintenance mode.
2. Ensure the official NVIDIA Enterprise Production Branch Driver (for servers) or the NVIDIA Studio Driver (for office endpoints) is integrated.
3. Stage the secure automation credentials inside the local Data Protection API (DPAPI) vault by running:

   cmdkey /generic:"VirusTC-API-Handshake" /user:"SystemNode" /pass:"YOUR_SECURE_CLUSTER_TOKEN"

4. Execute scripts/Citrix-TUF-Endpoint-Optimization.ps1 to apply the required registry overrides and system performance profiles. Shut down and seal the image.

Step 3: Deploy the Central API Listener & Dashboard

1. Place dashboards/VirusTC-Compliance-Dashboard.html within the document root folder (C:\inetpub\wwwroot\) of an internal administrative IIS web server.
2. Configure a secure SSL binding certificate on port 8443.
3. Launch scripts/VirusTC-API-Listener.ps1 as a continuous background system service on that host to process inbound node tracking payloads.

Step 4: Configure Active Directory Group Policy (GPO)

1. Link a new Group Policy Object to the Organizational Unit (OU) containing your diagnostic workstations.
2. Route to: Computer Configuration ➔ Policies ➔ Windows Settings ➔ Scripts (Startup/Shutdown).
3. Add scripts/VirusTC-SelfHealing-Auditor.ps1 as a Computer Startup Script. It must run in the computer context (NT AUTHORITY\SYSTEM) to ensure it has the administrative privileges required to repair network policies and rewrite HKLM registry paths.
4. Schedule scripts/VirusTC-Monthly-Report-Generator.ps1 and scripts/VirusTC-Monthly-Email-Subscription.ps1 on an administrative management host to deliver regular performance metrics directly to the hospital's Chief Information Officer (CIO).

---

🔒 Version Control and Verification Policy

All updates to the scripts or configurations in this repository must comply with the docs/GIT-Version-Control-Strategy.md protocol. Direct commits to the main branch are blocked.

Changes must be developed in a short-lived feature/* branch, clear the automated testing pipeline (.github/workflows/psscriptanalyzer-validation.yml), and undergo verification in an isolated staging lab environment before a pull request can be merged into production.

---

Disclaimer: This framework is an optimization architecture designed for integration by certified Healthcare IT and Clinical Systems Engineering teams. Verify all rendering states in a test lab before modifying live diagnostic software.