Muse 30B Glimmer
on NVIDIA RTX Pro 6000 Blackwell · TensorSharp
Summary
User benchmarks TensorSharp against llama.cpp on an RTX PRO 6000 Blackwell, with plain text generation decode at about 35 tok/s at 60 prompt tokens.
The user also tests DFlash speculative decoding and 2x RTX PRO 4000 Blackwell 24GB with tensor parallelism.