Yes, computes are carried out in FP16 (so there is no compute efficiency gains, might be latency reductions due to memory-bandwidth saving). These savings are n

Yes, computes are carried out in FP16 (so there is no compute efficiency gains, might be latency reductions due to memory-bandwidth saving). These savings are not realized yet because no custom kernels introduced yet.