Hey, it seems, if h264 input video contains frames with slices, there is no difference in decoding latency, nor throughput for nvdec when compared to video with non-sliced frames. How is it possible? Software decoders seem to heavily gain from sliced frames (if sliced-based threading is enabled).
