The microprocessor that was supposed to kill x86 but couldn't

Technology · Published 19 October 2026 Swipe more like this →

In short: In 1994, Intel and HP joined forces to create a processor so advanced it was predicted to supplant every existing architecture. They aimed to replace both RISC and CISC designs with a revolutionary system, yet this ambitious dream eventually stalled due to massive manufacturing hurdles and disappointing early performance.

Intel and HP launched the IA-64 architecture (later known as Itanium) with the goal of creating a dominant instruction set architecture (ISA) for all general-purpose computing. The project was born from HP's research into Explicitly Parallel Instruction Computing (EPIC), a successor to Very Long Instruction Word (VLIW) computing. Unlike conventional processors that handle instruction scheduling in hardware at runtime, EPIC required compilers to find and bundle independent instructions, theoretically allowing for much higher performance and better efficiency of silicon area and electricity. This ambitious project aimed to supplant both RISC and CISC architectures. HP researchers first developed the EPIC concept to balance silicon efficiency with general-purpose flexibility. In 1993, an internal competition saw the EPIC team beat the RISC team by over double the simulated performance.

However, the development of the first chip, codenamed Merced, was plagued by issues. A team of 500 designers—many of whom were inexperienced recent graduates—struggled with a design that was far too large for its intended manufacturing process. Complexity was scaled back by 1996, involving a smaller x86 component and a 96 KB L2 cache limit. These delays forced a shift in strategy; the focus moved to the second generation, McKinley, leading to the The project faced significant hurdles because the initial floorplan was unexpectedly massive. To meet size targets, the team eventually switched to a 180 nm process instead of the planned 250 nm. These complications meant the first chip became a mere development vehicle for the ecosystem.

wait for McKinley" This mantra was popularized by Fred Pollack, who claimed the second chip would double the first's speed. Meanwhile, HP extended its PA-RISC line to the PA-8900 to bridge the gap. By 2001, the architecture finally launched but struggled with poor x86 emulation and software maintenance.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

• How a 1970s architecture became the backbone of modern global computing
• How a failed pursuit of raw speed birthed modern computing efficiency
• The microprocessor that was too good for its own success
• The 1991 Apple-IBM-Motorola alliance aimed to break the Intel-Microsoft monopoly on personal computing.
• The architecture that powered the Macintosh, Sega Genesis, and Space Shuttle
• Java was first named Oak, after the tree outside its creator's office
• How a 1980s hotel meeting created the universal language for optical discs
• The Java Virtual Machine is an abstract computer that defies simple hardware categorization.
• Is a degree still necessary in the age of AI and bootcamps?
• JavaScript got its name as a marketing ploy, not from Java
• Born in a Finnish university, this text-based protocol once connected millions across the globe.
• How one computer scientist helped build the architecture powering almost every modern digital device.

More in Computing and information technology · More in Technology · Latest · A–Z · All topics