Give one example of each of the following. Be detailed and expand the answers with specific examples and detailed explanations whenever needed.

a.An applications where VLIW is superior in performance to superscalar.

[Answer Hint]â€“ This is a subset of the answer expected. Please use this concept and present a detailed explanation using graphs, charts, and at least 1 example to support your presentation.

VLIW hardware is simpler than superscalar hardware since the logic for detecting dependencies does not have to be included. Superscalar processors have a limited issue width due to the dependency checking complexity. Since VLIWs do not have this complex detection hardware, more room is available for wider issue, more functional units, and higher speeds. If an application has wide parallelism, the VLIW is better. Compiler methods can search a larger window for ILP.

b. An example where a VLIW is superior in performance to a vector processor.

[Answer hint] â€“ This is a subset of the answer expected. Please use this concept and present a detailed explanation using graphs, charts, and at least 1 example to support your presentation.

VLIW is better when different operations occur in parallel instead of the same operation on multiple data elements.

c. An example where a superscalar is superior in performance to a VLIW.

Superscalars have a superior performance when dependencies are better determined at run time or if a VLIW compiler is not powerful enough to identify the same dependencies statically. It is also Superior to running legacy code that would have to be transformed by binary-binary translation or code-morphing (Transmeta) on a VLIW.

-------------------------------------------------------------

d. An example where a vector processor is superior in performance to a VLIW.

Vector is better when the same operation occurs on multiple data elements (SIMD) instead of having different operations occur in parallel. This occurs quite a bit in scientific computing.

Give one example of each of the following. Be detailed and expand...