IA-64 Article

Already as of this date has multiple of the leading server and workstation suppliers announced their plans to incorporate IA-64 processors in their servers and workstations. Some of these companies are Microsoft, Adobe and IBM.

Today's processor architectures compilers create sequential machine code that attempts to imply parallelism to the hardware. Then must the processor's hardware reinterpret the machine code and try to identify opportunites for paralell execution. Paralell execution is the mainly key to higher performance and the better the processor can do this the faster it is. This process is inefficient not only because the hardware doesn't always interpret the compiler's intentions correctly, but also because it uses valuable die area that could be better used to do real work ( like executing instructions).

The IA-64 architecture's use of explicit parallelism enables far more effective parallel execution of software instructions. In the new IA-64 architecture model, the compiler analyzes and explicitly identifies parallelism in the software at compile time. This allows the most optimal structuring of the machine code to deliver maximum Instruction-Level Parallelism before the processor executes it rather than potentially wasting valuable processor cycles at run time. The result is significantly improved processor utilization. Also, there is no wasting of precious die area for the hardware reorder engine used in out-of-order RISC processors.

to the next page...