fp8_grouped_active_cap

Function fp8_grouped_active_cap 

Source
pub fn fp8_grouped_active_cap(
    num_tokens: u32,
    top_k: u32,
    num_experts: u32,
) -> u32
Expand description

2026-09-25: Cap on active experts, which sizes the grouped grids’ Y extent: num_tokens * top_k rows can reach at most that many distinct experts. It does not depend on the routing, so a captured graph stays valid for every routing.