Comments

• (Quote) 87.76.179.1

  in Post some YABS bench here Comment by Neoon June 9
• 3.25€/m Madrid # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## Yet-Another-Bench-Script ## v2026-05-11 ## https://github.com/masonr/yet-another-bench-script ## ## ## ## ## ## …

  in Post some YABS bench here Comment by Neoon June 8
• 1.28€/m Macao # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## Yet-Another-Bench-Script ## v2026-05-11 ## https://github.com/masonr/yet-another-bench-script ## ## ## ## ## ## #…

  in Post some YABS bench here Comment by Neoon June 8
• 1.99 Netgrid.Host Barcelona # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## Yet-Another-Bench-Script ## v2026-05-11 ## https://github.com/masonr/yet-another-bench-script ## ##…

  in Post some YABS bench here Comment by Neoon June 8
• 0.80€/m Dubaaaaai VM root@basic-vm-dubai-01:~# curl -sL https://yabs.sh | bash# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## Yet-Another-Bench-Script ## v2026-05-11 ## https:…

  in Post some YABS bench here Comment by Neoon June 8
• (Quote) No, they can differ, so does performance.

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon June 8
• (Quote) Again, I am not using Qwen3-VL

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon June 8
• https://www.reddit.com/r/LocalLLaMA/comments/1tzbcyp/llamacpp_gemma4_mtp_support_merged/ (Image)

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 7
• EnLighten mahNuts

  in nLighten confiscated hardware of MIRhosting and its colocation customers Comment by Neoon June 4
• I can fit Q4 on my 32gigs, Q6 is too tight. Q6 fits easily on the 64gig idle dedi though. Google just released a new gemma model, might be worth trying, just 12B so fits 16GB. https://www.reddit.com/r/LocalLLaMA/comments/1tvtn6m/googlegemma412b_hug…

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 4
• (Quote) Interesting, my local setup gives me about the same as the KS-LE-B.

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 3
• (Quote) what hardware? Q6 on KS-LE-B was 9/t for me.

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 3
• (Quote) Actually, ik has a webinterface now. github.com/ikawrakow/ik_llama.cpp its pretty barebones and it has way less features than the original llama.cpp. BUT, I get about 9t/s stable on the KS-LE-B vs the 6t/s on the llama.cpp one with Qwen 3.6…

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 3
• When South West?

  in TierHive General thread | Discuss, Updates, questions | LATEST: Static Hosting and PRICE UPDATE!!!! Comment by Neoon June 2
• Look mom, we are on TV https://point.free/blog/gemma-4-on-a-2016-xeon/

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 1
• (Quote) No idea, currently not using Qwen3 VL, just Qwen 3.5/6 35B

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon June 1
• (Quote) Yea, you juse load the correct model in llama.cpp with vision enabled. You click upload and it processes your picture. The upload button is disabled until the model is loaded though.

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 1
• (Quote) just input, image generation on CPU is painful.

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon June 1
• (Quote) Vision on Qwen is amaze, other models suck ass. The higher the res, the longer it takes obviously.

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon May 31
• Now we are cooking. (Image)

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon May 31
• https://www.reddit.com/r/LocalLLaMA/comments/1tluma3/llamacpp_server_have_builtin_native_tools_exec/ You can enable build-in tools, without anything extra. (Image)

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon May 31
• If you enable MTP, make sure the model supports it. You can find Qwen 3.5/3.6 here: https://huggingface.co/unsloth/Qwen3.6-35B-A3B-MTP-GGUF https://huggingface.co/unsloth/Qwen3.5-35B-A3B-MTP-GGUF

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon May 31
• To enable vision support, MPT and the recommended parameters, you can provide a config file for llama.cpp to load. I copied the original of the subreddit, this is mine currently. https://pastebin.com/raw/ZLP5t0fc You just have to provide --models-p…

  in How to (Ab)Use your KS-LE-B for LLM Models Comment by Neoon May 31
• https://www.reddit.com/r/LocalLLaMA/comments/1tr7hzw/psa/ (Image)

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon May 31
• (Quote) https://lowendspirit.com/discussion/10471/how-to-ab-use-your-ks-le-b-for-llm-models

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon May 31
• (Quote) Sadly I don't have a 16GB Card, hence I can't fit that model into memory. Skill issue, then run it on CPU, you can actually for once read it all.

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon May 31
• (Quote) Indeed and for such a KS-LE-B with 64gig DDR4 is purfect.

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon May 31
• (Quote) No, just get a RX 6000, it only has 96GB of VRAM yes, but its fast as fuck boy.

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon May 30
• BS, if you wanna run something with 10t/s get a KS-LE-B. Unified Memory isn't as fast as a RTX 6000 Blackwell. If you wanna run things fast, get a RTX 6000. You will be able to run these models on Unified Memory for sure, but it ain't as …

  in Run local LLMs on GPU or on "AI" Mini PC with unified memory? Comment by Neoon May 30
• wanna hear a funny story from software dev? Our lead dev. that is with the project for years, decided to dig for security issues, one day. The day he digged, he found 3. One of them was pretty ugly.

  in WHMCS critical vulnerability - all versions - patch now Comment by Neoon May 18
• (Quote) idk, if they don't have a test, to check if an addonID doesn't match a client, money isnt worth spend.

  in WHMCS critical vulnerability - all versions - patch now Comment by Neoon May 18
• big oof "when a client area user submits a request that includes an addonId parameter, WHMCS fails to verify whether that addon actually belongs to the requesting account. Swap in a different user’s addonId, and you walk straight into their se…

  in WHMCS critical vulnerability - all versions - patch now Comment by Neoon May 18
• (Quote) Kernel module loading for containers has been disabled on nodes where its possible. The rest still have to be upgraded to 6.x until we can go and disable it. As soon as news hits, these modules will be blacklisted asap anyway by me. If you …

  in microLXC Public Test Comment by Neoon May 13
• bro, do I really have to whitelist fkn modules, kurwa.

  in another day another LPE exploit - Fragnesia Comment by Neoon May 13
• Regarding the failed deployments today and the unavailability of the console / shell. A internal cert expired, it has been renewed but this caused deployments to fail and the console / shell to be unavailable for a few hours today.

  in microLXC Public Test Comment by Neoon May 13
More Comments