AI Briefing
KO

GHOST v2.1 Released

·2026.04.19 05:07

Key point

GHOST v2.1 adds native Windows support, simplifying running AI models on AMD GPUs.

Details

The open-source environment manager GHOST has been updated to v2.1, providing native Windows support.

To run AI models on AMD hardware, it automatically injects ZLUDA and ROCm layers, and is designed to run directly in PowerShell without needing Linux or WSL2.

Key features include the following.

  • Automatic hardware mapping: Scans the system and customizes the RDNA architecture settings needed for CUDA compatibility
  • Multi-GPU prioritization: Prioritizes detection of high-performance discrete GPUs over laptop integrated graphics
  • Stability improvements: Logic to prevent overlapping executions and process lifecycle management
  • Waiting screen TUI: A terminal UI that plays DOOM and music while models are loading
  • Safe mode fallback: Unsupported hardware automatically switches to an RDNA2 baseline

A repository link has also been made available.

open-sourcewindowsrocmai-inferenceamd-gpudevtoolszluda
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditOpen-Weight CLM Released as Jev Alternative, Agent Decision-Making 4x FasterPyTorchKRTrillion Labs Releases GravityOCR, an OCR Model Applying Self-Speculative DecodingHacker NewsJevper Released: A Jev Classification Wrapper for OpenAI-Compatible Models with Automatic Logprobs DetectionGeekNewsRootless Linux Sandbox 'Drop' Released for Isolating AI Agents and Third-Party ProgramsPyTorchKRPostgreSQL Extension pgjev Released, Supporting Natural Language Filtering Without EmbeddingsTLDR AIJared Palmer Releases 'Kev', a Lightweight Decision-Making Model Based on Qwen3.5