Qwen 3.6 27B flags/settings in llama.cpp

I run the following on a 5090 and have been okay with its performance, it does most things somewhere 80-100 t/s, though that can slow down at full 262k context - more like 40 t/s at times. I use it primarily in appdev tasks. This just barely fits in the 5090, no vision, with very very little room to

中文全文翻译正在补齐，可先查看下方英文原文。

I run the following on a 5090 and have been okay with its performance, it does most things somewhere 80-100 t/s, though that can slow down at full 262k context - more like 40 t/s at times. I use it primarily in appdev tasks. This just barely fits in the 5090, no vision, with very very little room to