2x CMP 170HX 64GB: GLM-5.3-Flash at 384K context / ~90 tok/s (EXL3, HBM-first setup) + Qwen3.8 comparison
英文标题：2x CMP 170HX 64GB: GLM-5.3-Flash at 384K context / ~90 tok/s (EXL3, HBM-first setup) + Qwen3.8 comparison
I've been tinkering with GLM-5.3-Flash on two 64GB CMP 170HX cards for a while, and the setup is finally stable enough that I figured I'd share it. I also compared it against the Qwen3.8-Flash-Next setup I've been using on the same machine: AWQ INT4 + FP8 PLE on vLLM. Besides PP/TG benchmarks,
打开原文相关 3 · 机会 1
中文
I've been tinkering with GLM-5.3-Flash on two 64GB CMP 170HX cards for a while, and the setup is finally stable enough that I figured I'd share it. I also compared it against the Qwen3.8-Flash-Next setup I've been using on the same machine: AWQ INT4 + FP8 PLE on vLLM. Besides PP/TG benchmarks,
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
I've been tinkering with GLM-5.3-Flash on two 64GB CMP 170HX cards for a while, and the setup is finally stable enough that I figured I'd share it. I also compared it against the Qwen3.8-Flash-Next setup I've been using on the same machine: AWQ INT4 + FP8 PLE on vLLM. Besides PP/TG benchmarks,