https://hgpu.org/?p=4446
Design and implementation of a time-division multiplexing scan architecture using serializer and deserializer in GPU chips