PulseAugur
EN
LIVE 16:23:23

Homelabber runs massive 2.8T-parameter AI model on CPU using Optane memory

By PulseAugur Editorial · [2 sources] ·

A Reddit user has successfully run Moonshot AI's 2.8-trillion-parameter Kimi K3 AI model using only a CPU and second-generation Intel Optane Persistent Memory modules. This setup, which eschews GPUs, achieved a speed of 0.85 tokens per second with 1-bit quantization, though it generated significant heat requiring active cooling. The user opted for Optane due to its lower cost and higher capacity compared to DRAM, making it a viable, albeit slow, option for running extremely large models in a home lab environment. AI

IMPACT Demonstrates a low-cost, albeit slow, method for running extremely large AI models locally, potentially enabling more experimentation outside of data centers.

RANK_REASON This is a user-driven application of existing hardware to run a large AI model, not a release from a frontier lab or a significant industry shift.

Read on Mastodon — mastodon.social →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

Homelabber runs massive 2.8T-parameter AI model on CPU using Optane memory

How we ranked this

Signal score
8 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
This is a user-driven application of existing hardware to run a large AI model, not a release from a frontier lab or a significant industry shift.
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
infra, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [2]

  1. Tom's Hardware TIER_1 English(EN) · Shane Downing ·

    Homelabber uses second-gen Optane memory sticks to run a 2.8-trillion-parameter AI model on CPU alone

    A Reddit homelab user ran Moonshot AI’s 2.8-trillion-parameter Kimi K3 CPU-only on Optane PMem 200 at 0.85 tokens/s with a 1-bit quant.

  2. Mastodon — mastodon.social TIER_1 English(EN) · hermes_vector ·

    Homelabber Runs 2.8-Trillion-Parameter AI Model on CPU With Optane Memory A Reddit user ran Moonshot AI's Kimi K3 model on a 40-core Xeon CPU using second-gen O

    Homelabber Runs 2.8-Trillion-Parameter AI Model on CPU With Optane Memory A Reddit user ran Moonshot AI's Kimi K3 model on a 40-core Xeon CPU using second-gen Optane PMem 200 memory, achieving 0.85 tokens per second with 1-bit quantization. https:// hermes-vector.buzz/posts/homel…

    LINKS hermes-vector.buzz/…/homelabber-runs-2-8-…

RELATED ENTITIES

RELATED TOPICS