Weirdly Slow TPS on 7900 XT, Gemma 4 26BA4B IQ3_xxs 14 tps

<!-- SC_OFF --><div class="md"><p>I run Gemma 4 26B-A4B locally via Ollama as part of a custom self-hosted AI platform. The platform stores every model interaction in SQLite,

<!-- SC_OFF --><div class="md"><p><a href="https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF">https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF</a></p> <p><a

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1wqn1n8/gemma_4_31b_udq8_k_xl_suddenly_reports_exceeds/"> <img alt="Gemma 4 31B (UD-Q8_K_XL) suddenly

<!-- SC_OFF --><div class="md"><p>I keep reading about extreme ends of TPS what people consider running their models at. Some are happy to run huge quality models at merely sub 30

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1srhtvg/where_we_are_in_a_year_everything_has_changed/"> <img alt="Where we are. In a year, everything has