Ubuntu 24.04 so slower than my Win11 for Qwen3.5-35B

<!-- SC_OFF --><div class="md"><p>Got Qwen 3.5 27B to 1,103,941 tok/s on 12 nodes with 96 B200 GPUs.</p> <p>At that rate you process 50,000 insurance policy documents in hours

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1t4uj9h/mtp_on_strix_halo_with_llamacpp_pr_22673/"> <img alt="MTP on strix halo with llama.cpp (PR #22673)"

<!-- SC_OFF --><div class="md"><p>Every output token the model generates requires reading the entire KV cache from High Bandwidth Memory.</p> <p>Here's what that actually looks

<!-- SC_OFF --><div class="md"><p>So i recently updated LMstudio after a long pause and updated my llama.cpp runtimes too.. i was shocked.. i thought maybe something like

<!-- SC_OFF --><div class="md"><p>That foodtruck bench post showing deepseek v4 matching gpt-5.2 at 17x cheaper got me thinking. if frontier cloud models are that overpriced for

<!-- SC_OFF --><div class="md"><p>Running Llama-3-70B with long contexts and hitting HBM limits constantly. 128K context burns approx 40GB just for KV cache — half the H100 gone

<!-- SC_OFF --><div class="md"><p>Hi,</p> <p>I taught Computer Science for 30 years in a French School of Electrical Engineering, Computer Science Department.</p> <p>I recently

<!-- SC_OFF --><div class="md"><p>I'm genuinely unsure if I'm doing something wrong, or if the model is just worse, despite the community enthusiasm. I tend to use it with pi