Hackernews posts about RTX 3090

1. RTX 5080 and RTX 3090 Setup: 80 Tok/s on Qwen 3.6 27B Q8 (imil.net)
   292 points by iMil 24 days ago | 108 comments
2. Story of How Im Running an Unlimited $6/Month AI Provider on 4x RTX 3090s
   8 points by yolo-auto 23 days ago | 3 comments
3. Qwen 3.6 93B with MTP on 2×RTX 3090 NVLink=187 tokens/SEC,LLM lost bleat-a-thon (github.com)
   4 points by devilfileprong 23 days ago | discuss
4. Nvidia resurrects older graphics cards as RAM demands impact tech prices (mashable.com)
   4 points by ripe 7 days ago | discuss
5. LLM from scratch, part 28 – training a base model from scratch on an RTX 3090 (www.gilesthomas.com)
   540 points by gpjt 7 months ago | 121 comments
6. Show HN: Llama 3.1 70B on a single RTX 3090 via NVMe-to-GPU bypassing the CPU (github.com)
   395 points by xaskasdf 5 months ago | 101 comments
7. We got 207 tok/s with Qwen3.5-27B on an RTX 3090 (github.com)
   165 points by GreenGames 3 months ago | 52 comments
8. Benchmarks show even an old Nvidia RTX 3090 is enough to serve LLMs to thousands (www.theregister.com)
   45 points by mfiguiere almost 2 years ago | 11 comments
9. Counter-Strike's Dust II runs purely within a neural network on an RTX 3090 (www.tomshardware.com)
   26 points by bpierre over 1 year ago | 3 comments
10. Dual RTX 5060 Ti 16GB vs. RTX 3090 for Local LLMs (www.hardware-corner.net)
    14 points by pietrushnic about 1 year ago | 5 comments
11. Show HN: OS Megakernel that match M5 Max Tok/w at 2x the Throughput on RTX 3090 (github.com)
    6 points by GreenGames 3 months ago | 1 comments
12. Qwen3.6-35B-A3B speculative decoding is net-negative on RTX 3090 (github.com)
    5 points by thc1006 3 months ago | 2 comments
13. PFlash: 10x prefill speedup over llama.cpp at 128K on a RTX 3090 (github.com)
    3 points by GreenGames 2 months ago | 1 comments
14. Upgrading Fans with a Custom Shroud on a RTX3090 – Goodbye Fan Noise (boilingsteam.com)
    3 points by ekianjo 8 months ago | 1 comments
15. Rent compute and run Ollama on an RTX 3090 – in one place (app.hyperbolic.xyz)
    3 points by hyperbolic_labs almost 2 years ago | 1 comments
16. I swept all 667 layer duplication configs for Qwen3-4B on 1x RTX 3090 (austinsnerdythings.com)
    3 points by auspiv 3 months ago | discuss
17. Upgrading Fans with a Custom Shroud on a RTX3090 For A Silent Experience (boilingsteam.com)
    3 points by ekianjo 8 months ago | discuss
18. Fine-Tuning Nvidia's GR00T 2B Robotic Model on a RTX 3090 (twitter.com)
    3 points by nprobably over 1 year ago | discuss
19. Club-3090 Recipes for serving QWEN3.6 27B locally on RTX 3090s (github.com)
    1 points by lastdong 2 months ago | discuss
20. Show HN: Federation of robots collaboratively train an object manipulation model (github.com)
    7 points by jafermarq over 1 year ago | discuss
21. Show HN: Open Access Qwen3.6-35B-A3B-UD-Q5_K_M with TurboQuant
    4 points by freakynit 3 months ago | 2 comments
22. Show HN: ZigNet: How I Built an MCP Server for Zig in 1.5 Days (fulgidus.github.io)
    4 points by Fulgidus 8 months ago | discuss
23. Show HN: I built a 2nd-order PyTorch optimizer for LLMs that runs on 16GB GPUs
    2 points by dnosoz 2 months ago | 4 comments
24. Show HN: Accidentally built churn detection AI (I don't code) (github.com)
    2 points by willster787 11 months ago | 2 comments
25. Show HN: kulyk-uk-en and kulyk-en-uk (huggingface.co)
    2 points by yehors 12 months ago | 1 comments
26. Show HN: Deep learning framework from scratch, trains GPT-2 in 3 days (github.com)
    2 points by netwrt almost 2 years ago | discuss
27. 70M vectors searched in 48ms on a single consumer GPU –results you won't believe
    1 points by EffCompute 4 months ago | 4 comments
28. Show HN: Localvoxtral – Local real-time dictation on macOS with streaming STT (github.com)
    1 points by T0mSIlver 4 months ago | 2 comments
29. AluminatiAI – per-job GPU cost tracking (Nvidia-smi shows watts, not dollars)
    1 points by AluminatiAi 4 months ago | discuss
30. Show HN: I over-engineered a home security camera that uses an LLM and talks (github.com)
    1 points by thecal 4 months ago | discuss
31. Show HN: Luna Agent – Custom AI agent in ~2300 lines of Python, no frameworks (nonatofabio.github.io)
    1 points by nonatofabio 4 months ago | discuss
32. Show HN: QLoRA fine-tuning in .zse INT4 format by ZSE
    1 points by zyoralabs 4 months ago | discuss
33. Show HN: GLM-Image Online – 16B AR+Diffusion model for accurate text (glmimage.online)
    1 points by hugh1st 6 months ago | discuss
34. 48GB RTX 8000 vs. 3090s for Local LLMs (www.hardware-corner.net)
    5 points by pietrushnic about 1 year ago | discuss