2015: Moore’s “Law” prevails for 50 years

Projection drove advances in semiconductor and magnetic storage

 

 

 

In April 1965, Electronics magazine published an article by Gordon E. Moore, Director of R&D at Fairchild Semiconductor that estimated the growth in the complexity of integrated circuits over the next ten years. He forecast that “by 1975 economics may dictate squeezing as many as 65,000 components on a single silicon chip.” This required a doubling every 12 months. In 1975, Moore reported that advances in photolithography, wafer size, process technology, and “circuit and device cleverness,” had allowed his projection to be realized.

On revisiting his projection with new data at that time, Moore modified the future rate of increase in the number of transistors on a chip to “a doubling every two years, rather than every year.” Originally intended to encourage his customers to adopt the most advanced technology in their new computer designs, his prediction emerged as a self-fulfilling prophecy. Dubbed Moore’s “Law,” semiconductor companies perceived it as a benchmark they must adhere to in order to stay ahead of the competition.

Even as Moore commented in 2003 that “No exponential change continues forever,” his revised prediction has prevailed. From 1971 (Intel 4004, 2,300 transistors) to 2015 (Nvidia Titan, 8,000,000,000), the number of transistors on the largest commercial microprocessor (MPU) devices increased at a compound annual growth (CAGR) rate of 41%. From 2,000 bits/sq. in. in 1957 (IBM RAMAC) to 1 trillion in 2015, commercial hard disk-drive (HDD) products have increased in areal density at a comparable CAGR to that of Moore’s Law for semiconductor devices. For HDDs, where operation depends on mechanical movement of the rotating disk and head track actuator, the improvement in speed performance however has greatly lagged the above CAGR values. These mechanical limitations have contributed to making SSDs attractive as HDD replacements in high performance applications. Efforts to produce ever faster and denser SSD solutions includes research into approaches as diverse as Phase Change Memory (PCM), ReRAM/RRAM, and carbon nanotubes.