Split mode "graph" for Cohere2 (#1061)

* This works and TG is descent, but PP is low

* Better

* Apply f_logit_scale before mul mat with output tensor

* This is better for PP: 600 t/s -> 700 t/s

* To not lose this again

* WIP

* Equal split

* WIP

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>

src/llama.cpp

@@ -1729,6 +1729,7 @@ static bool is_model_split_supported(const llama_model & model) {
         LLM_ARCH_QWEN3MOE,
         LLM_ARCH_GLM4_MOE,
         LLM_ARCH_MISTRAL3,
+        LLM_ARCH_COHERE2,
     };
     auto it =  k_supported.find(model.arch);
     return it != k_supported.end();