Guest viewpoint: a critical look at IA-64 (Massive resources, massive ILP, but can it deliver?)

Martin Hopkins

Abstract

Intel and HP have now released enough information for a preliminary evaluation of whether IA-64 is really the fundamental breakthrough in computer architecture that has been professed. In this author's opinion, it is not. This critique focuses on integer performance. Many important issues, such as floating point, system instructions, multimedia, and x86-compatibility mode are not covered, nor is this an evaluation if Itanium (Merced), McKinley, or any particular implementation. The intent is to look at the concept of EPIC (explicitly parallel instruction-set computing) as embodied in IA-64 and how it may work out in practice.