native PP256 and separate TG256

on AMD Strix Halo 128GB · llama.cpp / Vulkan RADV (original build unreported)

Oct 14, 2026

Throughput

31.1 t/s gen · 104.8 tokens/s prompt processing (prefill)Prompt processing measures input prefill speed; it does not tell you the time to first token.

Quant

UD-Q3_K_XL (GGUF)

Flash Attention

System RAM

128 GB

Summary

User reports Minimax M2.1 at 31.1 t/s generation and 104.8 t/s prompt processing on an AMD Strix Halo 128GB machine. Setup is llama-bench on Vulkan0 with the UD-Q3_K_XL quant, -ngl 999, -p 256 -n 256, -t 16, -r 3 and -fa 1. The native figures are PP256 104.80 +/- 7.95 and TG256 31.13 +/- 0.02; a separate toolbox run gave PP256 104.50 +/- 8.57 and TG256 33.07 +/- 0.05.