chickeyton
chickeyton
There are lots of protential algorithms for the Global Scheduler, the most important thing is providing a way to lookup the cache (i.e. `query_global_prefix_tree`), I suggest Mooncake provides a FullLookup...
> @chickeyton Could you write down the interface if FullLookup? sure, I put details in the description
/gemini summary
@YaoJiayi please review
May be there is somthing wrong in CI for the UT tests/v1/storage_backend/test_gds_backend.py ``` RuntimeError: cuFileHandleRegister failed (cuFile err=5020, cuda_err=0) ``` @Shaoting-Feng please help, thanks!
@YaoJiayi Please review again, the experiment of TTFT Routing shows some significant improvement and this PR is an essential feature for that