GLM-5.2 UD-IQ2_M runs at ~7.3 tokens per second on 4×RTX 3090s with 192GB DDR5 RAM using llama.cpp expert offload. Reducing quantization from IQ2 to IQ1 provided no speed gain, while increasing CPU threads from 6 to 12 improved performance by 22%. Decode is limited by CPU compute, not memory bandwidth, and the offloaded experts must be explicitly distributed across GPUs to avoid out-of-memory errors.
GLM-5.2 (744B, 2-bit) achieves 7.3 tok/s on 4×3090 with 192GB RAM
from English