Function rms_norm_short_row_eligible 

pub fn rms_norm_short_row_eligible(num_rows: u32, hidden_size: u32) -> bool

2026-09-25: Gate for rms_norm_warp_row: hidden_size <= 256 and even, and num_rows >= 1024. METRALE_RMS_NORM_WARP_ROW=0 turns it off; the variable is read once per process.