5. In 1991 Nick Tredennick said:

Question

“Superpipelining is a new and special term meaning pipelining. The prefix is attached to increase the probability of funding for research proposals. There is no theoretical basis distinguishing superpipelining from pipelining. Etymology of the term is probably similar to the derivation of the now?common terms, methodology and functionality as pompous substitutes for method and function. The novelty of the term superpipelining lies in its
reliance on a prefix rather than a suffix for the pompous extension of the root word.”

What do you think of this comment?

Answer 1

Nick Tredennick is a well?known writer and commentator on computer architecture. He tends to take a rather cynical view of the computer world. His comment is that superpipelining is a pretentious extension of pipelining. Essentially, he is correct. There are no rules about the number of stages in a pipelining or the rate at which they are clocked. However, for processor technology, implementation, instruction set, and workload, there will be an optimum pipeline length. The concept of superpipelining is basically increasing the number of pipeline stages if there are stages that are able to perform their role in much less than the allocated time slot. Suppose a pipeline has slots that take 2 ns, 4 ns, 8 ns, and 3 ns. The clock would have to have an 8 ns cycle time
to accommodate the longest state. The latency would be 4 × 8 ns = 32 ns and the instruction rate one per 8 ns.
If the 8 ns was divided into two 4 ns units, the pipeline would be 2, 4, 4, 4, 3 ns (five stages). The latency would
be 17 ns and the instruction rate would be one every 4ns. This is hardly a radical notion.