pub const FP8_QUANT_MIN_CTAS_PER_SM: u32 = 2;Expand description
CTAs per SM the twin insists on before it will take the launch.
Two, not one, and the same value as ssm_ba_gates_hopper’s
MIN_CTAS_PER_SM for the same reason: at exactly one CTA per SM a trailing
partial wave IS the kernel.