}Synchronous Design
}Latency of a pipeline
}Measured in terms of # of clock cycles
}Throughput
}Typically measured in terms of results per second
}Inverse of Clock Cycle Time for systems that generate a result
each cycle
}Asynchronous Pipelined Design (i.e., using pipelined
handshaking)
}Latency
}Time between tokens consumed at inputs and generated
outputs
}Inputs tokens spread apart to avoid congestion slowing down
results
}Cycle Time
}Taken as long-term average of time between successive output
tokens
}Throughput
}Results (tokens) per second - inverse of cycle time
}Data-dependent delays
}Block-level delays and data-flow may be data-dependent
}