YOU ARE VIEWING ONE ITEM FROM THE AICRIER FEED

LM Studio adds granular multi-GPU distribution controls

AICrier tracks AI developer news across Product Hunt, GitHub, Hacker News, YouTube, X, arXiv, and more. This page keeps the article you opened front and center while giving you a path into the live feed.

// WHAT AICRIER DOES

7+

TRACKED FEEDS

24/7

SCRAPED FEED

Short summaries, external links, screenshots, relevance scoring, tags, and featured picks for AI builders.

LM Studio adds granular multi-GPU distribution controls
OPEN LINK ↗
// 52d agoPRODUCT UPDATE

LM Studio adds granular multi-GPU distribution controls

LM Studio has introduced advanced GPU management tools, allowing users to move beyond basic automatic splitting to "Priority Order" and "Split Evenly" strategies. These controls are critical for developers optimizing performance across heterogeneous hardware or managing concurrent workloads like LLM inference and image generation.

// ANALYSIS

The addition of "Priority Order" mode effectively solves the bottleneck issues inherent in mixed-GPU rigs by allowing the fastest hardware to take the lead.

  • Users can now drag-and-drop GPUs to define a "fill-first" priority, ensuring the highest-performing card handles the bulk of the model layers.
  • A dedicated GPU Controls panel (Ctrl/Cmd + Shift + H) provides granular toggles for disabling specific cards or limiting offload to dedicated VRAM.
  • The "Priority Order" strategy is a superior alternative to fixed percentage splits, as it dynamically adapts to the specific VRAM limits of the hardware.
  • These updates solidify LM Studio's position as a power-user favorite for local LLM experimentation on consumer hardware.
// TAGS
lm-studiogpuinferencellmmlopsself-hosted

DISCOVERED

52d ago

2026-04-05

PUBLISHED

52d ago

2026-04-04

RELEVANCE

8/ 10

AUTHOR

CuriousEvilWeasel